Cancer Cell, Volume 22
Tumor Type-Dependent Function
of the Par3 Polarity Protein
in Skin Tumorigenesis
Sandra Iden, Wilhelmina E. van Riel, Ronny Schäfer, Ji-Ying Song, Tomonori Hirose,
Shigeo Ohno, and John G. Collard
Inventory of Supplemental Information
- Figure S1, related to Figure 1.
- Figure S2, related to Figure 2.
- Figure S3, related to Figure 3.
- Figure S4, related to Figure 6.
- Table S1, related to Figure 6.
Supplemental Experimental Procedures
Figure S1. Related to Figure 1. Examples of tumor histology in skin tumor experiments
with K14Cre+;Par3flox/flox mice. Paraffin-embedded tissue sections. (A) Typical papilloma found
in control mice. Note the exophytic morphology. Bar, 500um. (B) Basosquamous carcinoma
found in one K14Cre+;Par3flox/flox mouse. Bar, 500um. (C) Squamous cell carcinoma observed in
a K14Cre+;Par3wt/wt mouse. Bar, 500um. (D) Typical keratoacanthoma found in with
K14Cre+;Par3flox/flox mice. Bar, 500um. (E) Local invasion of papilloma into stroma cells. Bar,
100um. (F) Invasive cells of squamous cell carcinoma. Bar, 100um. (G) Local invasion of
basosquamous carcinoma cells into stroma. Bar, 100um. (H) Example for local invasion of
tumor cells into stroma as observed in a subset of Par3 KO keratoacanthomas. Bar, 100um. (I)
Immunohistochemical analysis of keratoacanthoma. Cortical staining of the adherens junction
protein E-cadherin is retained in the majority of tumor cells (left panel). Keratin-1, a marker for
differentiated keratinocytes, is expressed in cells of the tumor center but absent in the most
basal cell layers (middle panel). Tumor cells are usually negative for the mesenchymal marker
vimentin, indicating absence of EMT (right panel). ROI: higher magnification of the indicated
areas. Bars, 1mm.
Figure S2. Related to Figure 2. Loss of Par3 results in altered survival and apoptotic
signaling and cell motility. (A) Immunofluorescence staining of Par3 KO keratinocytes
transiently transfected with myc-PAR3 (green) 6hrs after calcium switch. Straight ZO-1 signal
(red) at intercellular contacts indicate formation of TJs. Bar: 20um. (B) Akt activity in murine
epidermal WT and Par3 KO keratinocytes. Non-treated or RasV12-expressing keratinocytes
were serum-starved for 3 hours before lysis. Total cell lysates were blotted for Par3 and P-Akt
(Ser473), stripped and reblotted for total Akt as loading control. Bottom: Quantification of above
(n=3) (mean ± SEM). *: p<0.05, **: p<0.01. (C) Top panel: Control or RasV12-expressing WT
and Par3 KO cells were cultured at NC for 36hrs, serum starved and treated with TPA or EGF.
Apoptosis was indicated by Caspase-3 cleavage, cyclin D1 expression served as read-out for
proliferation signaling. Numbers refer to relative cyclin D1 signal compared to α-tubulin (loading
control). Bottom panel: Quantification of relative cyclin D1 expression in WT and Par3 KO cells
upon various treatments (mean ± SD). n.s.: not significant, **: p<0.01. (D) Western Blot analysis
to detect apoptosis in WT and Par3 KO keratinocytes subjected to a calcium switch (2mM Ca2+)
36hrs before serum starvation (-GF, 8hrs). Total cell lysates were prepared from adherent and
floating cells and combined. Apoptosis was analyzed by Parp1 cleavage. *: full-length Parp1
(116kDa), >: cleaved Parp1 fragment (86 kDa) upon apoptotic signaling. Numbers indicate ratio
of the cleaved Parp1 fragment compared to loading control. α-tub, α-tubulin. GF: growth factor
presence (+) or depletion (-). (E) Western Blot analysis of Caspase-9 cleavage to assess
activation of the intrinsic apoptosis pathway. Asterisk: full-length pro-Caspase-9 (top: short
exposure, bottom: long exposure); arrow: fully processed 29kDa fragment of Caspase-9. Rac1
serves as loading control. Quantification is shown in Figure 2I. (F) Bax activation assay.
Immunofluorescence-based detection of conformational change and mitochondrial translocation
of Bax, indicative for Bax activation and subsequent activation of the mitochondria-dependent
intrinsic apoptosis pathway. Left: Costaining of active Bax (6A7 mAb) with mitochondria
(Tom20). Asterisks highlight cells positive for the active Bax conformation. Bars: 10um. Middle
panel: Co-immunostaining of active Bax (6A7, green) and total Bax (red). Bars: 10um. Right
panel: Quantification of Bax activation as cells positive for mitochondria-based 6A7 signal per
total cell count (mean ± SD). ***: p<0.001. (G) Left panel: Scratch wound assays of Par3 WT
and KO keratinocytes at 2mM Ca2+-containing medium, directly after scratch (t=0) or 24hrs
post-wounding. Bars: 20um. Right panel: Quantification of left (n=4) (mean ± SEM). *: p<0.05,
**: p<0.005. (H) Western blot analysis of cell-cell contact protein expression in total lysates of
primary mouse keratinocytes infected with RasV12 and incubated at low calcium or for 36hrs at
indicated millimolar calcium concentrations. E-cad, E-cadherin; β-cat, β-catenin.
Figure S3. Related to Figure 3. aPKC depletion reduces ERK activity, and intercellular
contacts are required for Par3-mediated ERK activation. (A) Effect of siRNA-mediated
aPKC depletion on ERK activation upon TPA treatment of WT mouse keratinocytes. Right
panel: Quantification of ERK activity upon aPKC-depletion in control and TPA-treated
keratinocytes (n=3) (mean ± SD). (B) Western Blot analysis of apoptosis in control and aPKC-
depleted keratinocytes as measured by Parp1 cleavage. Asterisk: full-length Parp1, arrow:
cleaved Parp1 fragment. (C) Quantification of ERK activity upon inhibition of the aPKC-Par6
interaction by aurothiomalate (ATM). WT (left) and Par3 KO (right) keratinocytes were cultured
at normal calcium for 36hrs, growth-factor depleted for at least 3 hrs and incubated with 300uM
ATM 1 hr before addition of TPA (mean +SD). (D) Quantification of relative ERK
phosphorylation in WT and Par3 KO keratinocytes cultured at normal calcium (NC) conditions
where cells have firm cell-cell contacts or at low calcium (LC) in the absence of cell-cell
adhesions. Relative ERK activity was assessed by Western Blot against phosphorylated ERK
and comparison to the loading control (n≥4) (mean ± SD). n.s.: not significant, *: p<0.05, ***:
Figure S4. Related to Figure 6. Detection of Ras mutations in papilloma and
keratoacanthoma of DMBA/TPA-treated mice, and immunohistochemical analysis human
and mouse skin tumors. (A) Nested PCR, followed by restriction digest, to verify Ras
mutations in all tumors analyzed. Briefly, the PCR product of the Ras allele is 176bp. Mutations
in the codon 61 result in a novel XbaI site (CTA mutation, cleaved fragments of 91 and 85bp) or
a TaqI site (CGA mutation, cleaved fragments of 90 and 86bp), whereas the WT Ras allele
remains uncleaved. No mutation was detected in normal skin (*), one sample showed a CGA
mutation (**), all other tumor samples were positive for a CTA mutation. ns: normal skin, p:
papilloma, k: keratoacanthoma. Arrows indicate cleavage products. (B) PAR3 immunostaining
of human keratoacanthoma (KA), squamous cell carcinoma (SCC) and epidermis adjacent to
the tumors (cryosections). Inset: E-CADHERIN immunostaining reflecting cells of epithelial
origin. All images were recorded at identical conditions. Bars, 10um. Quantification of B is
shown in Figure 6H. (C) Immunofluorescent staining of PAR3 with the TJ-associated protein
ZO-1, the adherens junction protein E-CADHERIN and the polarity protein aPKC in healthy
human epidermis. Bars, 20um. (D) ZO-1 immunostaining in WT and Par3 KO mouse epidermis,
papilloma and KA tissue. All images were recorded at identical conditions. Bars, 10um.
Table S1, related to Figure 6.
% of mice
(tumor-bearing mice / total mice)
WT Par3 cKO
9% (1/11) (BSC)
Average tumor numbers
(tumors / total mice)
WT Par3 cKO
17,4 (261/15)c 2,73*** (30/11)c
0,09 (1/11) (BSC)
% of mice
(mice with invasive tumors / total mice)
Overview of histopathological examination of skin tumors. All individual tumors of at least
5mm (KA, SCC, big papillomas) have been collected for histopathological analysis. These
bigger tumors where cut in half, and the remaining part was used for biochemical and genetic
analysis. a: mouse without papilloma but with KA, b: mouse without KA but with papilloma. c: A
few small papilloma have been collected entirely for biochemical and genetic analyses.
Average numbers of papilloma based on macroscopic analyses can be found in Figure 1B,C,
and 6D. WT: K14Cre+/Par3wt/wt; Par3 cKO; K14Cre+/Par3flox/flox, BSC: basosquamous
carcinoma. **: p<0.01, ***: p<0.001.
SUPPLEMENTAL EXPERIMENTAL PROCEDURES
Conditional Par3 deletion in the epidermis
The loxP-flanked exon 3 of Par3 (Hirose et al., 2006) was detected by PCR with primers 5'-
AGGCTAGCCTGGGTGATTTGAGACC-3' and 5'-TTCCCTGAGGCC TGACACTCCAG TC-3',
the Cre cDNA, inserted in the Keratin14 locus, was detected by PCR using primers K14-Cre3
(CGATGCAACGAGTGATGAGGTTC) and K14-Cre5 (GCACGTTCACC GGCATCAAC).
Plasmids, transfection and viral transduction
For rescue experiments, pK-mycPAR3b, a common human PAR3A splice variant (accession
number PAR3b: AF467003; Gao et al., 2002), and aPKCl-CAAX (kind gift of Susanne Vorhagen
and Carien Niessen) was used. For retroviral transductions, pBabe puro-SV40LT, hit&run Cre
recombinase (HR-MMPCreGFP) (Silver and Livingston, 2001) and pLZRS-HA-HRasV12 have
been used. Plasmids were transiently transfected into 70-90% confluent keratinocyte
monolayers using X-tremeGENE HP DNA Transfection Reagent (Roche) according to the
manufacturer’s protocol. Retroviral vectors were tranfected into Phoenix cells, and virus
supernatants were used to transduce keratinocytes as described previously (Michiels et al.,
siRNA-mediated aPKC downregulation
To target both aPKC isoforms, subconfluent mouse keratinocytes have been transfected with
ON-TARGETplus SmartPools against aPKCzeta and lambda (Thermo Scientific, mouse PRKCI
#L-040822-00-0005 and mouse PRKCZ #L-040823-00-0005) using 50nM of each SmartPool, or
non-targeting control pool (Thermo Scientific, #D-001810-10-05), and X-tremeGENE siRNA
Reagent (Roche) according to the manufacturer’s protocol. Efficient downregulation of both
aPKC isoforms was observed between 30 and 60 hours post-transfection.
Detection of oncogenic Ras mutations
Genomic DNA was isolated from normal and tumor tissue using TRIzol® Reagent. Ras alleles
have been amplified by nested PCR, and codon 61 mutations were detected as previously
described (Malliri et al., 2002; Finch et al., 1996) by subsequent digest with restriction enzymes
XbaI and TaqI.
Activation of Caspase-9, Caspase-3 and Parp-1 was detected by Western blot analysis of their
cleaved, active fragments. Bax activation was analysed by immunodetection of conformational
change of Bax on starved control and Ras-infected keratinocytes cultured for 32h at 2mM
calcium. Cells were fixed with 4% PFA and permeabilized using 0,05% Saponin. Both the
conformational change and the mitochondrial translocation of Bax were determined.
Inhibition of aPKC-Par6 interaction
Confluent keratinocytes were growth factor-depleted for at least 3hrs, then incubated for 1 hr
with 300uM aurothiomalate (SIGMA #157201), and then treated for 2 hours with 0,68uM TPA in
the presence of inhibitor before lysis.
Scratch wound assays were performed on confluent WT and Par3 KO keratinocytes cultured at
2mM calcium-containing medium. Artificial wound closure was followed by time-lapse
Immunoblotting and immunoprecipitation
Total lysates of cultured cells were prepared in hot SDS lysis buffer (1% SDS, 10mM EDTA,
and protease inhibitor cocktail, SIGMA). For immunoprecipitation, keratinocytes were lysed in
buffer containing 50mM Tris-HCl pH 7.4, 150mM NaCl, 0.5% Triton X-100. Cell extracts were
cleared by centrifugation and pre-cleared with protein A-sepharose for 1hr at 4°C. IP-antibodies
were coupled to fresh protein A-sepharose beads, and immunoprecipitation out of pre-cleared
lysates was performed for 4hrs at 4°C. Immunocomplexes were collected, washed five times
and eluted from the beads using SDS sample buffer. Separation by SDS-PAGE (NuPAGE,
Invitrogen) and immunoblotting was performed according to standard procedures.
Living cells were visualized using a phase-contrast microscope (Axiovert 25, Carl Zeiss
MicroImaging, Inc.). Whole slides of H&E- or DAB-stained tissue sections were scanned at 40x
magnification using automated slide scanners (ScanScope XT, Aperio Technologies, Inc., or
Leica SCN400, Leica Microsystems, Germany), and analysed using corresponding viewer
software provided by the manufacturers (Aperio ImageScope, Leica SCNViewer), or Tissue IA
(SlidePath, Ireland). Images of immunofluorescent samples were captured using a conventional
confocal microscope (TCS SP2, Leica, Germany) or spinning disk confocal microscope (Perkin
Elmer), or alternatively an epifluorescence microscope (Olympus IX81), and analysis was
performed using Image J (NIH, USA), Volocity (Perkin Elmer), and Photoshop (Adobe).
Quantification of protein signals in Western Blot analyses
The band intensity of non-saturated Western Blot signals was determined using the Adobe
Photoshop histogram tool on original digital files of scanned Western Blots. Bar diagrams show
the relative protein or phosphorylation signals after normalization using loading controls.
Quantification of Par3 expression in tissue sections
Tissue sections of healthy skin and different tumor specimens were processed for
immunostaining as described above. Micrographs of Par3 signal were taken using the spinning
disk confocal microscope (Perkin Elmer) at non-saturated detection levels, and identical imaging
conditions among epidermal and tumor areas. The signal intensities in epidermis adjacent to the
tumors or within tumor areas were determined using the automated measurement tool of
Volocity software (Perkin Elmer).
Student's t test was performed for quantifications of cell proliferation, apoptosis, active ERK, Akt
and Bax intermediates, tumor numbers and protein expression using Microsoft Excel software.
The asterisks shown in graphs correspond to the p-values as stated in the figure legends.
Measures of pooled data are represented by mean and standard deviation, SD, or standard
error of the mean, SEM, as indicated in the figure legends.
Gao, L., Macara, I.G., Joberty, G. (2002). Multiple splice variants of Par3 and of a novel related
gene, Par3L, produce proteins with different binding properties. Gene. 294, 99-107.
Michiels,F., van der Kammen,R.A., Janssen,L., Nolan,G., and Collard,J.G. (2000). Expression
of Rho GTPases using retroviral vectors. Methods Enzymol. 325, 295-302.