Suppression of KCMF1 by constitutive high CD99 expression is involved in the migratory ability of Ewing's sarcoma cells.

Abstract

High CD99 expression levels and rearrangements of the EWS gene with ETS transcription factor genes characterize the Ewing's sarcoma family of tumors (ESFT). CD99 is a cell surface glycoprotein whose engagement has been implicated in cell proliferation as well as upregulation and transport of several transmembrane proteins in hematopoietic cells. In ESFT, antibody ligation of CD99 induces fast homotypic cell aggregation and cell death although its functional role in these processes remains largely unknown. Here, using an RNAi approach, we studied for the first time the consequences of modulated CD99 expression in six different ESFT cell lines, representing the most frequent variant forms of EWS gene rearrangement. CD99 suppression resulted in growth inhibition and reduced migration of ESFT cells. Among genes whose expression changes in response to CD99 modulation, the potassium-channel modulatory factor KCMF1 was consistently upregulated. In a series of 22 primary ESFT, KCMF1 expression levels inversely correlated with CD99 abundancy. Cells forced to express ectopic KCMF1 showed a similar reduction in migratory ability as CD99 silenced ESFT cells. Our results suggest that in ESFT, high CD99 expression levels contribute to the malignant properties of ESFT by promoting growth and migration of tumor cells and identify KCMF1 as a potential metastasis suppressor gene downregulated by high constitutive CD99 expression in ESFT.

Full text

Page 1

SHORT COMMUNICATION
Suppression of KCMF1 by constitutive high CD99 expression is involved in
the migratory ability of Ewing’s sarcoma cells
M Kreppel1, DNT Aryee1, K-L Schaefer2, G Amann3, R Kofler4, C Poremba2and H Kovar1
1Children’s Cancer Research Institute, Kinderspitalgasse, Vienna, Austria;2Institute of Pathology, Heinrich-Heine-University,
Moorenstrasse, Dusseldorf, Germany;3Institute of Pathology, Medical University, Waehringer Guertel, Vienna, Austria and
4Tyrolean Cancer Research Institute, Innrain, Innsbruck, Austria
High CD99 expression levels and rearrangements of the
EWS gene with ETS transcription factor genes char-
acterize the Ewing’s sarcoma family of tumors (ESFT).
CD99 is a cell surface glycoprotein whose engagement has
been implicated in cell proliferation as well as upregula-
tion and transport of several transmembrane proteins in
hematopoietic cells. In ESFT, antibody ligation of CD99
induces fast homotypic cell aggregation and cell death
although its functional role in these processes remains
largely unknown. Here, using an RNAi approach, we
studied for the first time the consequences of modulated
CD99 expression in six different ESFT cell lines,
representing the most frequent variant forms of EWS
gene rearrangement. CD99 suppression resulted in growth
inhibition and reduced migration of ESFT cells. Among
genes whose expression changes in response to CD99
modulation, the potassium-channel modulatory factor
KCMF1 was consistently upregulated. In a series of 22
primary ESFT, KCMF1 expression levels inversely
correlated with CD99 abundancy. Cells forced to express
ectopic KCMF1 showed a similar reduction in migratory
ability as CD99 silenced ESFT cells. Our results suggest
that in ESFT, high CD99 expression levels contribute to
the malignant properties of ESFT by promoting growth
and migration of tumor cells and identify KCMF1 as a
potential metastasis suppressor gene downregulated by
high constitutive CD99 expression in ESFT.
Oncogene advance online publication, 12 December 2005;
doi:10.1038/sj.onc.1209300
Keywords: Ewing’s
migration; ion channels
sarcoma; EWS-FL11;CD99;
The Ewing’s sarcoma family of tumors (ESFT) is a
group of primitive small-round-cell tumors of bone and
soft tissue that show an extremely aggressive clinical
course but with an as yet unknown histogenic origin.
Common to this class of tumors is the presence of a
EWS gene-rearrangement with FLI1, ERG or, in rare
cases, other ETS genes resulting in the expression of
potent chimeric transcription-factors with both activat-
ing and repressing properties (Arvand and Denny,
2001). There is some variability in the architecture of
these fusion-proteins that may be of biological impor-
tance. The most frequent rearrangements are EWS-FLI1
type 1 and 2 occuring in 51 and 27%, respectively, and
EWS-ERG in 10% of cases. A further hallmark of
ESFT is a consistent high-level expression of the cell-
surface glycoprotein CD99 (Kovar et al., 1990; Ambros
et al., 1991). Introduction of EWS-FLI1 into neuro-
blastoma or rhabdomyosarcoma cells turns on CD99
expression (Rorie et al., 2004; Hu-Lieskovan et al.,
2005) while silencing of EWS-FLI1 in ESFT cells does
not affect high-level CD99 expression (our unpu-
blished results). Thus, the functional relation between
EWS-FLI1 and CD99 in ESFT remains unclear.
CD99 is a ubiquitous, 32kDa transmembrane sialo-
glycoprotein that is encoded by the pseudoautosomal
MIC2 gene (Goodfellow et al., 1988). Among normal
tissues, CD99 has been shown to be highly expressed on
pancreatic islet cells, Leydig and Sertoli cells (Ambros
et al., 1991). In the hematopoietic system, the level of
CD99 expression decreases with the degree of differ-
entiation, with early T- and B-precursors showing the
highest CD99 positivity (Gelin et al., 1989; Dworzak
et al., 1994, 1999). Similarly, an inverse association of
CD99 with differentiation has been demonstrated
recently for osteoblasts (Bertaux et al., 2005). Con-
versely, downregulation of CD99 is a major requirement
for generating Hodgkin’s and Reed-Sternberg cells in
Hodgkin’s disease (Kim et al., 1998, 2000). With the
exception of lymphoblastic lymphoma, ESFT remain
the only class of tumors that most consistently expresses
CD99, a feature widely employed in the differential
diagnosis of ESFT among small-round-cell tumors of
childhood (Fellinger et al., 1991). CD99 shares no
structural homology with any known protein family and
putative rodent orthologs significantly diverge from
human CD99 (Park et al., 2005). Available functional
data on CD99 derive from triggering CD99 by agonistic
monoclonal antibodies. In hematopoietic cells, CD99
ligation has been implicated in induction of homotypic
aggregation, cell adhesion, migration (Bernard et al.,
1995; Hahn et al., 1997; Schenkel et al., 2002),
upregulation of TCR expression (Choi et al., 1998),
Received 31 August 2005; revised 24 October 2005; accepted 26 October
2005
Correspondence: Dr H Kovar, Children’s Cancer Research Institute,
St. Anna Kinderspital, Kinderspitalgasse 6, Vienna A1090, Austria.
E-mail: heinrich.kovar@ccri.univie.ac.at
Oncogene (2005), 1–6
& 2005 Nature Publishing Group All rights reserved 0950-9232/05 $30.00
www.nature.com/onc

Page 2

vesicular transport of MHC molecules (Sohn et al.,
2001), and costimulation to induce a Th1-type cytokine
production (Waclavicek et al., 1998). In immature
thymocytes and ESFT cells, CD99 antibodies induce
massive apoptosis (Bernard et al., 1997; Sohn et al., 1998)
as well as caspase-independent cell death (Cerisano
et al., 2004) and increase sensitivity to chemotherapeutic
agents (Scotlandi et al., 2000). Recent studies have
provided fragmentary information about the involve-
ment of protein kinases in CD99 triggering of intracel-
lular signal-transduction (Hahn et al., 2000; Kasinrerk
et al., 2000). In most tissues, CD99 is expressed in two
isoforms generated by alternative splicing. In the B
lymphoblastoid cell line IM-9, the short-isoform which
lacks the cytoplasmic domain, has been demonstrated to
act antagonistically to the major long-isoform in the
induction of cellular adhesion (Hahn et al., 1997). In
addition, expression of the major form in a CD99-
deficient Jurkat T cell line was sufficient to promote cell
adhesion, whereas coexpression of the two isoforms was
required to trigger T-cell death (Alberti et al., 2002). The
diversity of biological responses to antibody-ligation of
CD99 in the absence of mechanistic insights into CD99
function and the lack of rodent homologues that might
be knocked-out for genetic studies prompted us to
investigate the role of CD99 in ESFT by an RNA-
interference approach.
Four different siRNAs targeting different regions of
both CD99 mRNA isoforms were designed and tested
for silencing CD99 expression after two consecutive
transfections in the ESFT cell lines SK-N-MC, TC252
and A673 expressing type 1 EWS-FLI1, STA-ET-7.2
and RDES carrying a type 2 EWS-FLI1 gene-fusion,
and RM82 with a EWS-ERG gene rearrangement. As
monitored by flow-cytometry on day 5 after the first
transfection (Figure 1a), all four siRNAs modulated
CD99 expression in SK-N-MC to variable extents. The
most effective siRNA targeting nucleotides 249–267 of
Figure 1
transfected twice with CD99 specifc siRNAs as indicated on days 1 and 3 with a final siRNA concentration of 200nM using
Oligofectaminet (Invitrogen, Carlsbad, CA). SiRNA sequences (all with dTdT 30overhangs) were: cd99si1 50-ACCCAGUGCUGGG
GAUGAC-30; cd99si2 50-CCCUAGUUCCUCCGGUAGC-30; cd99si3 50-AUGCCAACGCAGAGCCAGC-30; cd99si4 50-CCCACC
CAAACCGAUGCCA-30; cd99si2m (mismatched control) 50-CCCUAGUUCAGACGGUAGC-30; The cells were collected on day 5
after the first transfection and analysed for CD99 expression by either (a) flow-cytometry on a FACS-Calibur (BD Biosciences, San
Jose, CA), or (b) immunoblotting or (c) RQ–PCR. Protein expression of CD99 was monitored by monoclonal antibody 12E7 with
tubulin as a loading-control. Specific primers and probes for RQ-PCR were: for CD99: sense-primer 50-TAGGAGATGCTGTTGTT
GATGGA-30, antisense-primer 50-GGATTTGGCATCGGTTTGG-30, probe 50-AAAATGACGACCCACGACCACCGAA-30; for
beta-2-microglobulin: sense-primer 50-TGAGTATGCCTGCCGTGTGA-30, antisense-primer 50-TGATGCTGCTTACATGTCTC
GAT-30, probe 50-CCATGTGACTTTGTCACAGCCCAAGATAGTT-30. Cycling parameters on the ABI 7700 or 7900 sequence
detection system: 2min at 501C, 10min at 951C, and 50 cycles of 15s at 951C and 60s at 601C. The beta-2-microglobulin values were
used for normalization. CD99-silencing is expressed as the fold-decrease deduced from the difference between Ct-values of cd99si2 and
cd99si2m transfected cells.
RNAi-mediated silencing of CD99 in ESFT cell lines. SK-N-MC, STA-ET-7.2, RDES, A673, TC252 and RM82 were
CD99 suppresses KCMF1
M Kreppel et al
2
Oncogene

Page 3

the CD99 mRNA (cd99si2) downregulated CD99
surface-expression by 85% in up to 75% of transfected
cells as compared to untreated cells and cells transfected
with a mismatched control siRNA (cd99si2m). This
result was confirmed by immunoblot analysis of total
cell extracts from all transfected cell lines (Figure 1b)
and, on the RNA level, by real-time quantitative PCR
(RQ-PCR) (Figure 1c). A four- (TC252) to 10-fold
(RDES) reduction in CD99 expression was achieved at
day 5 after the first transfection. Efficient silencing
persisted until day 10 when CD99 levels were still
suppressed by more than 50% (not shown).
To investigate the effect of CD99-silencing on
surfacedependent and independent cell proliferation,
equal numbers of cd99si2 or cd99si2m transfected or
untreated SK-N-MC cells were either seeded on plastics
or in soft-agar. In monolayer cultures, only a small but
highly reproducible growth difference was observed
three days after seeding (Figure 2a). Identical results
were obtained with the cell line STA-ET-7.2 (not
shown). No other obvious changes in cell morphology
and growth-patterns were observed under this condition
(Figure 2b). This result was confirmed by histological
examination ofparaffin-embedded thin sections
Figure 2
collected directly after the second transfection and equal numbers seeded on cell culture flasks to assess surface-dependent growth.
Vital (Trypan-blue negative) cells were manually counted in a Buerker-Tuerk chamber at 48 and 72h post-seeding. Mean-values and
s.d. from three independent experiments are shown. (b) Surface-dependent growth pattern of SK-N-MC and STA-ET-7.2 cells treated
as in (a) (?10 magnification). (c) Immunohistology of STA-ET-7.2 cells left untreated or treated with cd99si2 or cd99si2m and grown
as spheroid cultures for 3 days (day 7 after the first transfection). Spheroids were fixed with 4% paraformaldehyde, sedimented, and
embedded in 1.5% agarose to maintain morphology. Samples were subsequently embedded in paraffin and thin sections stained with
CD99 antibody 12E7 with hemalum counterstain. (d) [3H]thymidine-incorporation assay of cd99si2 and cd99si2m-treated SK-N-MC.
Cells were transferred to 96-well flat-bottom tissue-culture plates and incubated with 1mCi (3.7?104Bq) [3H]thymidine (ICN
Biomedicals, Costa Mesa, CA) per well for 18h. [3H]thymidine-incorporation was assessed by b scintillation counting. Thymidine-
incorporation in cd99si2-transfected cells was calculated from three independent determinations and is displayed relative to control
transfection with cd99si2m. (e) Anchorage-independent growth was determind by seeding of siRNA-transfected STA-ET-7.2 cells into
soft-agar on top of a 0.5% agarose-containing basal layer. Colonies were manually counted 5, 6 and 7 days postseeding. Results from a
representative experiment are shown. (f) Morphological analysis of EFST cell colonies growing in soft-agar. Colonies of cd99si2 and
cd99si2m-transfected STA-ET-7.2 cells were checked for morphological features 6 days postseeding (?10 magnification; details
displayed as inserts at ?32 magnification).
CD99-silencing reduces ESFT cell growth. (a) SK-N-MC cells transfected with cd99si2 and cd99si2m (negative control) were
CD99 suppresses KCMF1
M Kreppel et al
3
Oncogene

Page 4

obtained at day 7 from cd99si2 and cd99si2m trans-
fected STA-ET-7.2 cells grown as spheroid cultures for 4
days (Figure 2c). Except for the expected difference in
CD99 membrane staining decorating the cells contours,
immuno staining for mesenchymal (vimentin), epithelial
(cytokeratin mix), and neural (monoclonal S100, neuro-
filament, chromogranin A, synaptophysin, monoclonal
neuron-specific enolase) markers was negative except for
vimentin in both control and cd99si2 treated cells (data
not shown). Together with the apparent lack of neurite-
like cellular processes these results are consistent with
the absence of differentiation in response to CD99
silencing. No evidence for increased cell death was
obtained by propidium iodide and Annexin V staining
(data not shown), but thymidine-incorporation assays
suggested that, upon CD99 silencing, the observed
growth-reduction was due to decreased proliferation
(Figure 2d). When seeded into 0.35% soft-agar, CD99
suppressed STA-ET-7.2 cells showed a marked decrease
in colony numbers as compared to controls (Figure 2e).
Moreover, the morphology of colonies was different.
Until day 10 (day 6 after seeding), while untransfected
(not shown) and mismatch control transfected cells gave
rise to loose colonies with cells penetrating the soft-agar
occasionally attaching to the basal layer underneath,
cd99si2 transfected cells grew as clumpy compact
colonies which were always smaller than the controls
(Figure 2e). Afterwards, CD99-suppressed colonies
started to grow normally and to spread-out consistent
with the loss of RNAi-suppression. These results may be
considered in context with previous reports on homo-
typic aggregation and impaired cellular migration of
hematopoietic and ESFT cells when CD99 was ligated
to antibody (Bernard et al., 1995; Scotlandi et al., 2000;
Schenkel et al., 2002).
To directly test the effect of CD99-silencing on the
mobility of ESFT cells, SK-N-MC and STA-ET-7.2 cells
were transfected with cd99si2 and cd99si2m, or were left
untreated, and tested for their migration-potential in
transwell-chamber assays. Upon modulation of CD99, a
50% reduction of migrating cells was consistently
obtained in both cell lines 1 day after seeding (Figure 3).
It should be noted that cd99si2 has the potential of
modulating both CD99 isoforms. Thus, the observed
changes in proliferation and migration of ESFT cells
cannot be assigned to a shift in relative expression levels
of full-length and truncated CD99 isoforms but is due to
the general loss of CD99 expression. The combined data
indicate that CD99 plays a role in both ESFT growth
and migration.
Figure 3
ESFT cells. 5?105siRNA-transfected and untransfected SK-N-MC
and STA-ET-7.2 cells were seeded into the upper compartment of
transwell-chambers (Corning, MA) and were incubated for 24h at
371C. Cells that migrated through the filter to reach the lower
chamber were fixed with 2.5% glutaraldehyde, stained with
hemalaun and counted using an inverted-microscope. The number
of cd99si2 and untransfected cells that have passed through the
filter is presented in percent relative to cd99si2m control transfected
cells.
CD99-silencing effect on the in vitro migratory ability of
Table 1Expression-profiling of CD99-suppressed cells
Gene name CelllineAverage fold change UniGeneGenBank
Genes upregulated upon CD99 silencing
Potassium channel modulatory factor 1
Placental growth factor, vascular endothelial growth factor-related protein
F-box protein 9
DnaJ (Hsp40) homolog, subfamily A, member 2
Myeloid cell leukemia sequence 1 (BCL2-related)
Proenkephalin
SK-N-MCy 1; STA-ET-7.2y 2; TC252y 3
1; 2; 3
1; 2; 3
2; 3
1; 3
1; 2
1; 2
2.5
2.0
1.9
1.8
1.7
1.7
Hs.345694
Hs.252820
Hs.216653
Hs.368078
Hs.532826
Hs.339831
AI743396
BC001422
AL137520
AI697792
BF981280
BC010527
Genes downregulated upon CD99 silencing
START domain containing 4, sterol regulated
BCL2-like 2
Insulin induced gene 1
RAB GTPase activating protein 1-like
Kinesin family member 21B
Myeloid/lymphoid or mixed-lineage leukemia (trithorax homolog, Drosophila)
Early B-cell factor
Homeodomain interacting protein kinase 3
SK-N-MCy 1; STA-ET-7.2y 2; TC252y 3
1; 2
2; 3
1; 2
1; 2
1; 2
1; 2
1; 2
1; 2
?1.9
?1.8
?1.8
?1.7
?1.7
?1.7
?1.7
?1.7
Hs.93842
Hs.410026
Hs.520819
Hs.495391
Hs.169182
Hs.258855
Hs.308048
Hs.201918
AA628398
BC021198
BG292233
AB019490
NM_017596
Z69744
W73890
AF305239
RNA extracted at day 5 post-transfection of cell lines TC252, SK-N-MC, and STA-ET-7.2 with cd99si2 or cd99si2m and tested for efficient CD99-
silencing was subjected to comparative gene-expression profiling-analysis on Affymetrix U133A (one experiment) and U133 Plus 2.0 microarrays
(three experiments). Only genes displaying a more than 1.6-fold mean expression difference in the comparative analysis of cd99si2 and cd99si2m-
transfected cells in at least two of the three cell lines are shown and were considered as putative CD99 downstream genes.
CD99 suppresses KCMF1
M Kreppel et al
4
Oncogene

Page 5

In order to gain insight into the mechanisms by which
CD99 regulates ESFT cell growth and migration, we
aimed at defining changes in gene-expression patterns
associated with modulated CD99 expression thus
determining the end point of CD99-signalling. SK-N-
MC, STA-ET-7.2 and TC252 cells were transiently
transfected twice consecutively with cd99si2 and, for
comparison, cd99si2m, in four independent experiments
and subjected to expression-profiling by microarray
analysis. Using a cutoff value of 1.6-fold, we found 37
probe sets representing 14 genes which were modulated
by CD99-silencing in at least two of the three tested cell
lines (Table 1). Only two genes, the potassium-channel
modulatory factor 1 (KCMF1) and placental growth
factor (PGF), showed consistent upregulation upon
RNAi-mediated CD99-silencing in all three cell lines.
To validate KCMF1 as a consistent downstream gene of
CD99 in ESFT, we performed RQ-PCR analyses on 6
ESFT cell lines with different EWS gene-rearrangements
5 days after the first (2 days after the second)
transfection. Figure 4a clearly indicates that KCMF1
was consistently induced up to six-fold by transient
CD99 silencing independent of the type of diagnostic
EWSgene-rearrangement.
result with all four siRNAs to CD99, as demonstrated
for SK-N-MCcells,confirmed the specificity of
the effect. These observations suggest that CD99, by
an unknown mechanism, generally represses KCMF1
expression in ESFT.
In addition, we performed Affymetrix GeneChip-
based gene expression patterning on a series of 22
primary ESFT (16 localized and six metastatic). Results
obtained for KCMF1 and CD99 indicated a statistically
significant inverse correlation between the expression
levels of these two genes (Pearson correlation coefficient
r¼?0.558, P¼0.007) consistent with the experimen-
tally observed CD99 dependent modulation of KCMF1
(Figure 4b).
To test, whether KCMF1 contributes to the observed
changes in the migratory ability of ESFT cells, KCMF1
cDNA was cloned from CD99-silenced SK-N-MC cells
with a C-terminal His-tag into a CMV promoter-based
expression vector and was stably introduced into the cell
lines SK-N-MC and STA-ET-7.2. When control (empty
vector transfected) and KCMF1 expressing polyclonal
cell populations were tested in transwell-chamber
assays, a marked difference in the migratory ability of
ESFT cells was observed (Figure 4c). KCMF1-trans-
fected cells showed a consistent reduction in migration
similar to ESFT cells with silenced CD99. Similar results
were obtained with the cell line RDES (not shown).
Taken together, these results imply that constitutive
CD99 expression supports ESFT cell migration by
suppression of KCMF1 expression in the absence of
additional stimulatory CD99 interactions. Since migra-
tory ability is an essential component of tumor-
dissemination andESFT
metastatic, our results identify KCMF1 as a CD99-
regulated putative metastasis suppressor gene. Future
studies will have to test this hypothesis in a xenograft
mouse model.
Reproducibilityofthis
are intrinsicallyhighly
Figure 4
upregulation of KCMF1 and reduced ESFT cell-migration. (a)
RNAi-mediated silencing of CD99 upregulates KCMF1 RNA
levels. Cell lines SK-N-MC, STA-ET-7.2, RDES, A673, TC252 and
RM-82 were transfected with the indicated siRNAs and, for
control with cd99si2m and analysed on day 5 by RQ–PCR for
CD99, KCMF1 and control beta-2-microglobulin expression as
described for Figure 1c. Primers and probe for KCMF1 analysis
were: sense-primer 50-CTTGAACACAGAGCCCCTAGAGA-30,
antisense-primer 50-CGGCCAGGGTGAAACATTC-30, probe 50-
GATGAATCGAGTGGTGTTCGACATGTACG-30. Mean va-
lues and s.d. of relative CD99 and KCMF1 expression-changes in
suppressed versus cd99si2m control-transfected cells from three
independent determinations performed in triplicate (except for
cd99si1, 3, and 4, and A673 and RM82 – single experiments each)
are shown. (b) Affymetrix GeneChip results for KCMF1 (probe set
217938_s_at) and CD99 (probe set 201029_s_at) expression for 22
primary ESFT at diagnosis (16 localized, 6 metastatic). (c) Effect of
KCMF1 overexpression on the in vitro migratory ability of ESFT
cells. KCMF1 cDNA was amplified from cd99si2-transfected
SK-N-MC cells using primers 50-ATGTCCCGACATGAAGTT
GTCA-30and 50-AAGAGGAGGTGGTGGAGGCTC-30, cloned
into pcDNA3.1/V5-His&TOPOsTA (Invitrogen, Carlsbad, CA)
and transfected into SK-N-MC and STA-ET-7.2 cells. 5?105
empty vector or KCMF1 expression vector-transfected cells were
seeded into transwell-chambers and analysed for their migratory
ability as described in the legend to Figure 3. Expression of ectopic
KCMF1 was monitored in parallel on the immunoblot using
anti-His6-Peroxidase antibody (Roche, Mannheim).
Silencing of constitutive CD99 expression results in
CD99 suppresses KCMF1
M Kreppel et al
5
Oncogene