article

1. [CANCER RESEARCH 58. 2624-2632. June 15. I998|
Disruption of the Cyclin D/Cyclin-dependent Kinasc/1NK4/Ketinoblastoina Protein
Regulatory Pathway in Human Neuroblastoma1
John Easton,2 Tie Wei, Jill M. Lahti, and Vincent J. Kidd3
t'ni of Tumor Cell Biology. St. Jude Children 's Research Hospital, Memphis, Tennessee 38101
ABSTRACT
The pl6INK4" (MTS1) and pl8INK4c gene products are normal, and
highly expressed, in human neuroblastoma cell lines. The retinoblastoma
protein (pRb) was, nonetheless, phosphorylated and functional in these
cells. Such high levels of pl6INK'4"/pl8INK'u'should normally inhibit cyclin-
dependent kinase (CDK) 4 and 6 activities in cells containing functional
pKb, delaying cell cycle progression and growth. These neuroblastoma cell
lines express both CDK4 and CDK6 mRNA and protein, but only signif
icant CDK6 protein kinase activity was detected in this study. In addition,
CDK6 was not present in plfi'SKl<" immune complexes in cells with
significant kinase activity, although pl6INK4l> levels were high. Others
have shown that a specific mutation in the NH,-terminal region of the
CDK4 gene product can disrupt |>1(>'NK'" binding, thereby bypassing its
inhibitory activity. To determine whether mutation of the CDK6 gene, or
some other mechanism, is responsible for the CDK6 kinase activity in
these cell lines, several complementary analyses were performed. The
CDK6 gene from each cell line was examined for mutations that might
affect pl6INK4" binding, whereas pl6INK" add-back experiments were
performed with CDK6 immune complexes to assess pl6INK4°function. A
bona fide CDK6 mutation that disrupts pl6INK4a binding and prevents
inhibition of CDK6 protein kinase activity was identified in 1 of 17
neuroblasloma cell lines. The mechanism(s) responsible for disruption of
pl6INK4" inhibitory activity in the remaining cell lines is unknown, but
these results suggest that neuroblastoma cells may bypass the cell cycle
block imposed by constitutive expression of wild-type pl6INK4li in novel
ways.
INTRODUCTION
NB4 is a pediatrie cancer that accounts for approximately 10% of
all childhood tumors (1). At the molecular level. NB appears to
include at least three separate tumor types with different genetic and
biological characteristics, but all share a common neural crest cell
origin. NBs characterized by MYCN gene amplification and the loss of
human chromosome Ip36 have a particularly poor prognosis (2).
Despite extensive efforts to determine the functional significance of
MYCN gene amplification in these tumors, its role is not entirely clear
(1). However, considering the effects of c-myc expression on cell
growth, it is likely that N-wivr overexpression affects both cell cycle
and apoptotic signaling pathways (3-7). Surprisingly, unlike many
tumors, the mutation/deletion of common tumor suppressor genes
[p53. RBI. and MTS1 (pl6lNK4*)] is infrequent in NB (1, 8). Even so,
Received 12/9/97; accepted 4/20/98.
The costs of publication of this article were defrayed in pan by the payment of page
charges. This article must therefore he hereby marked advertisement in accordance with
18 U.S.C. Section 1734 solely to indicate this fact.
' This research was supported by NIH Grant CA 67938 (to V. J. K.). Cancer Center
Core Grant CA 21765 (to St. Jude Children's Research Hospital), and by the American
Lebanese Syrian Associated Charities (to V. J. K. and J. M. L.).
' Present address: Department of Molecular Pharmacology. St. Jude Children's Re
search Hospital. Memphis, TN 38101.
' To whom requests for reprints should be addressed, at Department of Tumor Cell
Biology. Si. Jude Children's Research Hospital. 332 North Lauderdale. Memphis. TN
38101. Phone: (901 ) 495-3597; Fax: (901) 495-2381.
4 The abbreviations used are: NB. neuroblastoma; CDK. cyclin-dependent kinase;
CKI, cyclin-dencndenl kinase inhibitor; pRh. retinoblastoma protein; GST. glutathione
.S'-lransferase; KMSA. eleclrophoretic mobility shift assay; SSCP. single-strand confor-
malional polymorphism; FISH, fluorescence in situ hybridization; 1VTT. in vitro tran
scription and translation; RT-PCR. reverse transcription-PCR.
NB cells exhibit unregulated growth characteristics commonly asso
ciated with tumors harboring mutations in these genes.
The cyclin D/CDK4 and cyclin D/CDK6 protein kinases function,
at least in part, by regulating the phosphorylation status of pRb during
the cell cycle (9). These changes in pRb phosphorylation regulate its
growth-suppressive functions by affecting its interactions with other
proteins (e.g., E2F; Ref. 10). The protein kinase activity of the cyclin
D/CDK4/6 complex is regulated by its association with other protein
kinases, phosphatases, and CKIs (9, 11). The cyclin D/CDK/INK/pRb
cell cycle pathway is subject to specific alterations during tumorigen-
esis, presumably due to its importance in response to mitogenic
stimulation (12-21). The genetic alterations that occur in this pathway
during oncogenesis appear to involve many of its components, includ
ing the D-type cyclins, CDKs, and CKIs. Furthermore, the differen
tiation of NB cells requires the inhibition of CDK4 and CDK2
activities accompanied by the loss of pRb phosphorylation, strongly
suggesting the requirement of cyclin D/CDK and cyclin E/CDK
complexes for NB cell proliferation (22, 23).
Accumulating data suggest that specific modifications in either the
level or function of the D-type cyclins (particularly cyclin Dl),
pl61NK4a (MrS/)/P15INK4b (MTS2), CDK4/6, and/or pRb are in
volved in many, if not all, tumors (19, 24). Most frequently, mutation/
deletion of either MTS1 or RBI occurs in many tumors (19). A
strategic link between alterations in cyclin Dl and/or CDK4 levels
(due to either gene disruption or gene amplification), mutations that
affect the ability of the MTS1/MTS2 gene products to function nor
mally, and mutations in RBI was made by Peters and colleagues (11,
13, 19, 25). This crucial proliferati ve pathway is also affected by
specific mutations within CDK4. impairing the ability of the CDK4
protein to properly interact with and be inhibited by pl6INK4;l (26). All
of these alterations ultimately result in the functional inactivation of
pRb, thereby preventing pRb from suppressing cell growth appropri
ately (19, 26).
Here we report that NB cell lines express very high levels of normal
pl6INK4il and/or pl8INK4c, yet these cells contain paradoxically high
levels of cyclin D/CDK.6 protein kinase activity and functional pRb.
A bona fide mutation in CDK6, which prevents pl6INK4a from prop
erly inhibiting CDK6 protein kinase activity, was observed in 1 of 17
of these cell lines. This is the first report of a CDK6 mutation in
human tumors that is similar in nature to a CDK4 mutation described
in melanoma (20, 26). Examination of pl6INK4il immune complexes
from all of the remaining cell lines with high CDK6 kinase activity
revealed that these complexes did not contain appreciable amounts of
CDK6, although significant amounts of both proteins are found in
these cells. Furthermore, exogenous p!6INK4'' added back to CDK6
complexes from these same cells was capable of inhibiting this protein
kinase activity in vitro. These results suggest that a novel mechanism
is responsible for pl6INK4il inactivation in these NB cell lines.
MATERIALS AND METHODS
CDK4 and CDK6 Kinase Assays and CDK6 Kinase/GST-pl6INK4a
Add-Back Assays. Cell lysate was obtained from cells growing in log phase
using previously reported procedures (27). For each sample. 12(X)/^g of cell
lysate were incubated with 1 ¡j,gof anti-CDK6 antibody or anti-CDK4 anti
body (both from Santa Cruz Biotechnology) in a total volume of 4(X)/j.1on ice
2624
Research.
on September 8, 2019. © 1998 American Association for Cancercancerres.aacrjournals.orgDownloaded from

2. CDK6 MUTATION IN NEUROBLASTOMA
for 2 h. Protein A agarose (Pierce Chemical Co.) was then added (25 /u.1of a
50% slurry in lysis buffer), and the samples were incubated at 4°Cfor 45 min.
The pelleted beads were washed five times with lysis buffer and two times with
1 ml of kinase buffer [50 mM HEPES (pH 8), 10 mM MgCl,, 0.1 mm NaV, 1
mM NaF, 10 m.MB-glycerophosphate, and 1 mM EDTA]. After the final wash,
2 fig of GST-pRb fusion protein and 10 jxCi of [-y-32P]ATP were added to the
beads, incubated at 30°Cfor 30 min, and pelleted, and the samples were
electrophoresed on 7.5% SDS-PAGE gels. The signals were quantified using
a Phosphorlmager (model 475F; Molecular Dynamics). For add-backs, the
samples were prepared as described for the kinase assays using 800 /j.g of cell
lysate. The protein A beads were resuspended in 400 fil of lysis buffer, and 5
iig of GST-pl6INK4a were added to one set of samples, whereas a duplicate set
of samples received an equivalent amount of BSA. All of the samples were
incubated at 4°Cfor 4.5 h. The samples were washed one additional time with
1 ml of lysis buffer prior to washing with the kinase buffer.
Western Blots. Cells were lysed in ice-cold radioimmunoprecipitation
assay buffer, sheared three times through an 18-gauge needle, and incubated on
ice for 45 min. The debris was pelleted, and the supernatant was collected. For
Western blots. 200 /j.g of protein were loaded onto either a 7.5% (pRb), 10%
(CDKs and cyclins), or 12% (pl6INK4a) SDS-PAGE gel. The samples were
then transferred to an lmmobilon-P membrane (Millipore) using a semidry
transfer apparatus (Millipore) per the manufacturer's instructions. After incu
bation with primary and secondary antibodies the membranes were washed and
developed with enhanced chemiluminescence (Amersham Corp.) per manu
facturer's instructions. The antibodies were obtained from the following sup
pliers: Santa Cruz Biotechnology, anti-pi6'NK4a, anti-CDK4, anti-CDK6, anti-
cyclin Dl. and anti-cyclin D2, and PharMingen. human anti-pRb.
EMSAs. Cells from one 10-cm dish were harvested and extracts were
prepared as described by Scholer et al. (28) using 100 ¡JLof microextraction
buffer [20 mM HEPES (pH 7.7). 450 mM NaCI, 0.2 mM EDTA. 0.5 mM DTT,
25% glycerol. 100 ng/ml phenylmethylsulfonyl fluoride. 1 /ig/ml leupeptin. 1
fig/ml aprotinin, and 1 jug/ml pepstatin A] quantitated by a Bio-Rad protein
assay (Bio-Rad). The E2F binding site probe was prepared using two oligo-
nucleotides: E2FA, 5 '-AATTCCTTGGCGGGA AAAAG AAGOG A-3 '. and
E2FB, 5'-AGCTTCCCTTCTTTTTCCCGCC AAGO 3'. The underlined se
quence contains the strong E2F binding site II found in the Myc gene promoter
(29). and the oligonucleotides were labeled and purified as described. For each
NB cell line analyzed, a set of binding reactions was prepared using 10 /j.g of
cell lysate and 1 ¿tl(20,000 cpm) of E2F probe in a total volume of 25 /¿I(28.
29). The samples were incubated for 2 min at 25°Cbefore addition of 3 /xl of
human anti-pRb antibody (PharMingen) to one tube. The mixed samples were
placed at 25°C.incubated for 17 min, loaded onto a 4% PAGE gel made with
0.25 X Tris-borate EDTA, and electrophoresed at 270 V in the cold (4°C)for
2.5 h. The gels were then dried and autoradiographed.
SSCP Analysis, Genomic PCR, PI Phage Isolation, FISH, and Northern
Blotting. Two hundred ng of genomic DNA from the NB cell lines were used
in each PCR reaction. The oligonucleotide primer sets and reactions conditions
were identical to those used by Hussussian et al. (30) and Zariwala et al. (31),
except for the following modifications: 3 /*Ci of [a-12P]dCTP (DuPont NEN
Research Products) were added to the PCR reaction. The electrophoresis was
performed at both 15 W for 4-8 h at room temperature and 6-8 W for 12 h
at 4°C.
For PCR of CDK6, 200 ng of genomic DNA were used per reaction. The
primer set consisted of 5'-CAGTACGAATGCGTGGCG-3' (CDK6F) and
5'-CTCCTCGCCGGTCTGCAC-3' (CDK6R). The reaction conditions for the
PCR were 50 mM KC1, 10 mM Tris-Cl (pH 9.0), 0.1% Triton X-100, 0.2 mM
each deoxynucleotide triphosphate, 1.5 mM MgCU. and 100 ng of each primer
in a 50-/il reaction volume. PCR amplification was performed in a Perkin-
Elmer Corp. model 480 thermal cycler as follows: denaturation at 95°Cfor 5
min, 30 cycles of a 15-s denaturation at 94°C,a 15-s annealing at 56°C,and
a 30-s extension at 72°C.A final 3-min extension at 72°Cwas performed in the
final cycle. PCR products were cloned into the PCRII vector using the TA
cloning kit (Invitrogen). Individual colonies were isolated and analyzed.
Twenty samples were pooled and sequenced using 35S Dideoxy Sequencing kit
(United States Biochemicals), electrophoresed on 6% polyacrylamide sequenc
ing gels, and autoradiographed. The same technique was also used for CDK4
analysis; the primer set for CDK4 consisted of the forward primer 5'-
CGATATGAGCCAGTGGCTGAAATTGGT-3' (CDK4F) and the reverse
primer 5'-CTCCTCCTCCATTGGGGACTCACACTC-3' (CDK4R). Human
plgiNK4c p| phage ciones were isolated with the following primers: 5'-
ACTGCTACTTAGAGGTGCTAATC-3' (HP 18-1) and 5'-TGCATCAG-
GCTAACAACCTCATT-3' (HP18-2). PCR reaction buffer contained 50 mM
KC1, 10 mM Tris-Cl (pH 9.0), 0.1% Triton X-100, 0.2 mM each deoxynucle
otide triphosphate, 1.5 mM MgCl2, and 100 ng of each primer in a 50-jxl
reaction volume. Denaturation at 95°Cfor 4 min was followed by 30 cycles of
a 1-min denaturation at 94°C,a 45-s annealing at 62°C,and a 2-min extension
at 72°C.The final product corresponded to a 300-bp genomic fragment from
the />/8INK4c gene. Two different pl81NK4t PI phage clones were isolated
(Genome Systems. St. Louis, MO) and were denoted hpl8-l and hpl8-2.
These PI phage clones were used to perform FISH analysis of NB cell lines as
described previously (27).
The probe used to detect CDK6 mRNA consisted of a 584-bp fragment of
human CDK6 (GenBank accession no. X66365) extending from nucleotides
161-734. The probe used to detect CDK4 mRNA consisted of a 1157-bp
fragment of human CDK4 (GenBank accession no. M14505) extending from
nucleotides 73-1230. All of the other probes consisted of gel-purified full-
length cDNA inserts.
RT-PCR for CDK6 and pl6INK4". The reverse transcription reaction was
performed using 5 /j.g of total RNA and superscript II reverse transcriptase
(Life Technologies. Inc.). A gene-specific primer made to the 3'-untranslated
region of CDK6 was used: 5'-TCCAGAAGACAGCAGCTG-3' (CDK6-RT).
The resulting reactions were used in a PCR template amplification with the
Expand Long PCR (Boehringer Mannheim) system, with the forward primer
5'-CTTGCCATGGAGAAGGACGGCCTGTG-3' (CDK6-F4) and the reverse
primer 5'-CGCGGATCCGCTAAGGCGGCTGCTGAGG-3' (CDK6-R3).
The nucleotide sequence for an Ncol restriction site was added onto the 5'-end
of the CDK6-F4 primer, and the sequence for a BamHl restriction site was
included at the 5'-end of the CKD6-R3 primer, to allow cloning into the p-SK
vector. The amplification was carried out in a Robocycler (model 960; Strat-
agene) under the following conditions: denaturation for 2 min at 94°C.fol
lowed by 40 cycles with a 45-s denaturation at 94°C,a 45-s annealing at 62°C,
and a 1.5-min extension at 68°C.For pl6INK4a, the primers and conditions
specified by others were used (30, 32). The nucleotides and excess primer were
then removed using Microcon 100 spin columns (Amicon); the PCR products
were digested sequentially with Ncol and BamHl, purified using the Qiagen
Quick PCR Purification Kit (Qiagen Corp.). and cloned into Ncol/BamHl-
digested p-SK vector. DNA from individual colonies was isolated from each
sample using Qiagen minicolumns (Qiagen). Coupled transcription/translation
reactions were performed using the TNT rabbit reticulocyte lysate system
(Promega Corp).
Construction of CDK6 Mutants. The plasmid construct pSK-glob-CDK6
was used as a template for double-stranded site-directed mutagenesis using the
Chameleon mutagenesis kit (Stratagene). Oligonucleotides were synthesized
that incorporated the observed mutations: 5'-GGCGCCTATGGGAGGGTGT-
TCAAGG-3' (NB 10/12) and 5'-GTGTTCAAGGCCCGCIACTT-GAA-
GAACGGAG 3' (NB14). The underlined nucleotides indicate the position at
which the mutation was introduced.
Production of GST Fusion Proteins and pl6INK4a and pI8INK4c Binding
Assays. Escherichia coli containing a GST-pl6INK4a construct (kindly pro
vided by D. Beach, Cold Spring Harbor Laboratory) and a GST-pRb (amino
acids 379-928 of the human pRb; kindly provided by Dr. Mark Ewen, Dana
Farber Cancer Institute) were obtained, and proteins were purified as described
previously (33,46). IVTT protein products for the wild-type and mutant CDK6
gene products were combined with 5 /¿gof GST-pl6INK4a or GST-pl8INK4c
bound to glutathione Sepharose 4B beads (Pharmacia Biotech) in 0.5 ml of
NP40 lysis buffer [150 mM NaCI, 1.0% NP40. and 50 mM Tris (pH 8.0)] at 4°C
for 2 h. The beads were then washed and resuspended in 20 /il of 2X SDS
sample loading buffer, boiled, and analyzed on a 10% SDS-PAGE protein gel.
After electrophoresis, the gels were placed in Amplify (Amersham), dried, and
autoradiographed.
Baculovirus Expression of wtCDKó, CDK6-D32Y, and Cyclin Dl in Sf9
Cells. The wtCDK6 and CDK6-D32Y cDNAs described above, as well as a
previously described human cyclin Dl cDNA (34), were cloned into the
pFastBack baculovirus (Life Technologies, Inc.). A high-tiler stock of the pure
recombinant viruses were obtained and used to infect insect Sf9 cells to obtain
cyclin D1/CDK6 complexes. The production of recombinant human wtCDKo.
CDK6-D32Y. and cyclin Dl proteins was determined by Western blotting with
the appropriate antibodies. Protein kinase activity was assayed using a GST-
2625
Research.
on September 8, 2019. © 1998 American Association for Cancercancerres.aacrjournals.orgDownloaded from

3. CDK6 MUTATION IN NEUROBLASTOMA
pRb fusion protein as described earlier. The ability of exogenously added
GST-pl6INK4a protein to inhibit this protein kinase activity was assayed as
described previously by others (26).
RESULTS
Expression of D-type Cyclins, CDK4/6, pl6INK4a, pl8INK4c, and
pRb in NB Cell Lines. DNA, RNA. and protein from 17 different
human NB cell lines (NB1-NB21; see Ref. 27 and references therein),
many with amplified MYCN genes (see Table 1), were examined with
probes corresponding to the D-type cyclins, cyclin E, CDK2, CDK4,
CDK6, P161NK4a,P18INK4c,p2iWaf'-ciP''sdi'-CaP2°,p27Kipl, and pRb,
all of which are key regulators of the cell cycle that might be altered
in response to increased levels of N-myc' (6). Cyclin E and CDK2 are
expressed at normal, invariable levels in 16 of 17 of the cell lines
examined; NB3 expressed relatively high levels of cyclin E [and
several smaller molecular weight isoforms corresponding to alterna
tively spliced transcripts (35); data not shown). On the basis of these
molecular analyses, as well as significant morphological differences
between NB3 and the other NB cell lines, we believe that NB3
represents an atypical NB cell line. The CKIs p2lWaf'-ciPl-sdil-CaP20
and p27Klpl are expressed at significant levels in only one of the NB
cell lines, NB21, which contains a CDK4 amplicon (Table 1). There
fore, neither of these CKIs was examined further. All but five of the
NB cell lines (NB1. NB2, NB3, NB10, and NB16) expressed very
high levels of pl6INK4a (Fig. 1). Conversely, pl81NK4c was expressed
at very high levels in four of five of these NB cell lines, but only one
of the other NB cell lines (NB20) expressed similarly high levels of
this CKI (Fig. 1). Intriguingly, although the level of pl6INK4a and/or
pjgiNK4c mRNA and protein was markedly elevated in 16 of 17 of the
NB cell lines examined, both hyper- and hypophosphorylated forms of
pRb were detected by Western blot analysis in 16 of 17 of the same
cell lines (Fig. 1). Cyclin Dl was expressed at moderate to high levels
(16 of 17 NB cell lines; cyclin D2 was expressed in all NB cell lines,
including a cyclin D1-negative cell line), whereas CDK4 (17 of 17)
and CDK6 (16 of 17) are expressed at relatively equivalent levels in
all of the NB cell lines (Fig. 1 and data not shown). As expected, the
NB21 cell line containing the CDK4 gene amplicon expressed signif
icantly higher levels of the corresponding CDK4 protein.
The presence of such high levels of pl6INK4a or pl8INK4c in cells
containing hyperphosphorylated pRb normally suggests that there are
functional mutations in the MTSI gene (19, 24). Otherwise, mutations
that either compromise pRb function or circumvent the ability of
p]6iNK4a to ¡nn¡b¡tD-type cycIin/CDK kinase complexes are ob
served (20, 26). In NB, the MTSI gene is almost always normal, with
rare reports of a mutation or deletion in this gene (8, 36). Similarly,
the gene encoding pl8'NK4c, which is localized to Ip32, is rarely
mutated or deleted in any cancer, and it is not altered in any of the NB
cell lines examined thus far (31). None of the 17 NB cell lines had
deletion or rearrangement of the p/SINK4c gene when examined by
FISH (data not shown). Examination of the INK genes by SSCP
analysis revealed no apparent abnormalities, with the exception of the
deletion of the MTSI gene in NB l and NB 10, which accounts for the
absence of protein in these cell lines (Fig. 1). Finally, DNA sequenc
ing of RT-PCR products corresponding to MTSI revealed occasional
polymorphisms in exon 3, all of which have been reported previously
(37), but no mutations in NB2-8 and NB 12-21 (data not shown).
These observations were somewhat surprising, especially given that
pRb was efficiently hyperphosphorylated in all but the NB3 cell line
(Fig. 1).
Functional Analysis of pRb. Because these studies demonstrated
that very high levels of normal pl6'NK4a and hyperphosphorylated
pRb are present in many of the NB cell lines examined (Fig. 1), we
decided to examine pRb function by EMSAs and pRb antibody
supershifts (Fig. 2). The oligonucleotide probe used in the EMSA
corresponds to a canonical E2F binding site; if pRb is functional, it
will sequester the E2F-related (E2F1, E2F2, E2F3, and DPI) tran
scription factors when hypophosphorylated (9). Thus, functional at
tributes of pRb can be directly assessed by use of a 32P-labeled E2F
oligonucleotide probe, which should bind at least a portion of these
E2F-related proteins when they are in complexes with pRb. Similarly,
an appropriate pRb antibody should either supershift the complex or
eliminate it entirely. As controls for these experiments, we chose two
cell lines that have been well studied, MOLT4 and C33A. The human
T-cell line, MOLT4, has functional pRb, active D-type cyclin/CDK
complexes, and nonfunctional pl6INK4a, whereas the human cervical
carcinoma epithelial cell line C33A contains nonfunctional pRb, nor
mal pl6INK4a, cyclin D2, and CDK4.
A pRb-associated E2F complex can clearly be seen by EMSA for
all of the NB cell lines except NB3, which contains an E2F complex
that migrates more slowly than those in the other NB cell lines (Fig.
2). These pRb-associated E2F complexes comigrate with the complex
in the MOLT4 cells (the positive control), and a pRb/E2F complex is
Table 1 Characteristics of NB cell lines used in this study"
Cell
line[plófNBINB2
+/-NB3
+/-NB4
+4-4-NB5
++NB6
4-4.MN8
++NBIONB12
++NB13
+ ++NB14
++NB16
+NB17
4-4-4-NB19
4-4-4-NB2Ü
++NB21
4-4-4-4-C33A
4-4-4-4.Molt4[pi
8]* [pRbl*[CDKof4-4-4-4-
Y4-4-4-4-4-
Y4-+
NY
4-Y
+Y
4-+/-
Y++
Y4.Y
+Y
++
Y+4-4-
Y4-+/-
Y++
N+4-4-4-
Y++
Y+ND
NND
Y +CDK6
£16 £16 P16
[DU* activity CDK41 CDK6'mut4-4-
4-4-4-Y4-4-4-
4-4-N+
+ + 4-N+
+ 4- 4-N4-4-4-
+ + 4-N4-4-4-4-
4-4- 4-/—N+
+ 4-4-N+
+4-4-Y+
4-4- +/-N+/—
+/- + +N4-4-
4-+ 4- 4-N+
+4-4-N+
+N+
+/- +N+
+N+
+ + + + + +/-N+
N-
+4- - YCDK6
mutNNNNNNNNNNYNNNNNNDNDpl8mutNNNNNNNNNNNNNNNNNDNDCDK4mutNNNNNNNNNNNNNNNNNNCDK4ampNNNNNNNNNNNNNNNYNDNDpRbmutNNYNNNNNNNNNNNNNYNMYCN2-6>202
a mut. mutation: amp. amplification: Y. yes; N, no; ND, not determined.
''Apparent concentrations from Fig. 1.
' Presence ( + ) or absence (—)of CDK4 or CDK6 in api6 immune complexes from Fig. 2.
'' From Ref. 27 and references therein; del. one alÃ-eledeleted: trans, one alÃ-eletranslocated: del/trans, one alÃ-eledeleted, one alÃ-eletranslocated.
2626
Research.
on September 8, 2019. © 1998 American Association for Cancercancerres.aacrjournals.orgDownloaded from

4. CDK6 MUTATION IN NEUROBLASTOMA
p16
Fig. 1. Protein levels of cyclin DI. CDK4.
CDK6, pl6'NK4a, pl8'NK4c, and pRb in NB cell
lines. Western blots of the various NB cell lines
(indicated above each lane), and the C33A and
MOLT4 controls were sequentially probed with
antibodies to the proteins (indicated to the left of
each blot).
cdk4
CQ CO CÜ CO
cdk6
o
a
g- in
Z Z
cyclin D1
pRb
not present in the C33A cells (the negative control). In all of the NB
cell lines, except NB3, the pRb/E2F complex is disrupted (Fig. 2)
when an antibody to pRb is incubated with cellular proteins and the
E2F oligonucleotide prior to EMS A analysis. This confirms that these
are pRb-associated E2F complexes. The more slowly migrating E2F
complex observed in the NB3 cell line may be due to its association
with the Rb-related pocket protein, p 130, because this band can be
shifted with an antibody to p 130 (data not shown). In conclusion,
these experiments strongly suggest that these NB cell lines not only
contain hypo- and hyperphosphorylated pRb (Fig. 1), but that this pRb
is also functional (Fig. 2).
Presence of CDK4 and/or CDK6 in pl61NK4a Immune Com
plexes. Given the high level of normal pl6INK4aand normal levels of
CDK4 and CDK6 found in many of the NB cell lines, pl61NK4a
immune complexes should also contain one or both of these CDK
subunits at levels consistent with their steady-state levels of expres
sion (as determined by Western blotting; Fig. 1). When —200/ng of
total cell lysate from each NB cell line were immunoprecipitated with
an excess of pl6INK4a-.specific polyclonal antiserum and these im
mune complexes were analyzed for the presence of CDK4 and/or
CDK6 by Western blotting, this was not the case (Fig. 3). A number
of these pl6INK4a immune complexes did not contain CDK6, and
Fig. 2. Functional analysis of pRb by
EMSA. Lysate from each NB cell line was
incubated with a "12P-labeIed oligonucleotide
containing the E2F binding site to detect E2F
complexed with pRb. The location of the
pRb-containing complexes, in the absence of
apRb (apRb—), is shown (pRb arrow, left).
The NB cell line used is indicated above each
set of lanes. The positive control is MOLT4,
in which the pRb complex can clearly be
seen. Specificity of the E2F-pRb complex
(pRb arrow, left) is demonstrated by its dis
appearance when the anti-pRb antibody
(apRb+) is added to the reaction. C33A
serves as a negative control. No E2F-pRb
complexes are observed in the C33A sam
ples, regardless of the presence (+ ) or ab
sence (-) of apRb.
i i
T Iß CO h-
Z Z Z Z i i
£ t 5! w 5 «
CD 00 ffl ¿D O CO
trRb -
(ree ^„„g.
2627
Research.
on September 8, 2019. © 1998 American Association for Cancercancerres.aacrjournals.orgDownloaded from

5. CDK6 MUTATION IN NEUROBLASTOMA
Fig. 3. Detection of CDK4 and/or CDK6 in
pl6'NK4" immune complexes (apio). pl6'NK4°,and
the cellular proteins associated with this protein,
were precipitated from 200 ^.g of each NB cell
lysale (indicated above each lane). These api6
complexes were prohed with a combination of
CDK4 and CDK6 human anlirahbit polyclonal sera.
Left, location of the CDK4 and CDK6 proteins. The
preimmune control consisted of affinity-purified
IgG from rabbit.
*- CM
m m
z z
Su) co f~
CU CO CO
Z Z Z Z
00 CD CD 00
z z z z
(O
£
en o *-i- CM CM
CO CD CD
*
" §
co Q
cdk6-
cdk4-
several did not contain CDK6 or CDK4. This suggested that several
of these NB cell lines might harbor mutations in CDK6 or CDK4.
which affected their ability to bind pl6INK4a.
CDK4 and CDK6 Protein Kinase Activities in NB Cells. It was
equally important to determine whether any of the NB cell lines had
significant CDK4/6 kinase activity, particularly cell lines containing
significant amounts of CDK4 and/or CDK6 in their p 16INK4aimmune
complexes (NB4, NB5, NB7, NB12, NB13, NB17, NB19, NB20, and
NB21 ). CDK4- and CDK6-specific antisera, generated to the COOH-
terminal regions of these proteins and capable of efficiently immuno-
precipitating these kinase activities (38), were used to isolate both of
these kinase complexes using ~ 1 mg of total cell lysate from loga
rithmically growing NB cells. A GST-Rb fusion protein containing
the A and B "pocket" region of pRb that is normally phosphorylated
by the D-type/CDK4/6, as well as cyclin E/CDK2, was used as the
substrate for these in vitro protein kinase assays (Fig. 4A). As ex
pected, negligible CDK4 protein kinase activity was detected in all of
the NB cell lines. Somewhat surprisingly, however, moderate to very
high levels of CDK6 protein kinase activity were present in many of
the NB cell lines. In all cases, CDK6 kinase activity was high in NB
cell lines in which CDK6 was not readily detected in a pl6INK4a
immune complex. Once again, one possible interpretation of these
results is that CDK6 has been altered, possibly by somatic mutation,
in a way that affects the ability of pl6INK4a to interact with and inhibit
this protein kinase.
To further confirm this possibility, an in vitro CDK6 protein kinase
assay, in which a superphysiological level of exogenous pl6INK4a is
added back to the aCDKo immune complex, was used to determine
the relative sensitivity of the CDK6 protein kinase activity to pl6INK4a
inhibition (Fig. 4ß).A similar approach has been used to assay
functional mutations found in both pl6INK4a and CDK4 (20, 26). To
determine the concentration of exogenously added GST-pl6INK4a
fusion protein required to efficiently inhibit wild-type CDK6 protein
kinase activity from MOLT4 cells (a positive control containing
CDK6 kinase activity but no functional pl6INK4a), increasing amounts
of GST-pl6INK4a, or a BSA control, were added to identical aCDKo
immune complexes from these lysates. Although 5 p.g of the BSA
protein had no significant effect on CDK6 kinase activity, only 0.5 p.g
(10-fold less) of the GST-pl6INK4a fusion protein effectively inhibited
the CDK6 protein kinase (Fig. 4ß).
On the basis of these results, we used 5 /ng of exogenously added
GST-pl6INK4a protein (> 100-fold molar excess) to determine
whether the CDK6 kinase activity from the various NB cell lines was
affected in a similar manner. CDK6 protein kinase activity in most of
these NB cell lines could, indeed, be effectively inhibited by the
exogenously added GST-pl6INK4a (Fig. 45). However, the NB14 and
NB16 cell lines continued to express significant CDK6 kinase activ
ity, even in the presence of this normally effective inhibitor. It is
important to note that a mutation in the CDK6 gene in these NB cell
lines would affect only one of the two alÃ-eles;therefore, a portion of
the CDK6 protein kinase activity in these cells should correspond to
the wild-type protein encoded by the second alÃ-ele.This would result
in pl6INK4a-insensitive and pl6INK4a-sensitive pools of the CDK6
protein. Under these circumstances, one would expect at least a 50%
reduction in the ability of the CDK6 kinase to phosphorylate the
GST-pRb substrate. In fact, the kinase activity in the NB14 and NB17
aCDK6 immune complexes containing exogenously added GST-
pi 6INK4awas inhibited —50%(Fig. 4B; this experiment was repeated
multiple times and quantitated by use of a Molecular Dynamics 475F
Phosphorlmager). On the basis of these results, DNA sequence anal
ysis of RT-PCR and genomic PCR products corresponding to the
CDK6 genes in NB14 and NB16 was performed to determine whether
a mutation was present.
It should also be pointed out that, paradoxically, a number of the
NB cell lines contained significant levels of CDK6 protein kinase
activity that could be efficiently inhibited by the exogenously added
GST-pl6INK4a, but apparently not by the normal, and highly ex
pressed, endogenous cellular pl6INK4a (e.g., NB2, NB6, NB8, NB10,
NB12, and NB17). Although these data suggested that these NB cell
lines did not contain mutations in CDK6 that would affect their ability
to bind pl6INK4a, they were included in both SSCP and RT-PCR/
genomic PCR analyses. It is always possible that CDK6 gene muta
tions distinct from those that affect pl61NKa binding are present in
these cells.
SSCP Analysis, Genomic PCR, and DNA Sequence Analysis to
Confirm CDK6 Gene Mutations. SSCP analysis of exon 1 of the
CDK6 gene, which contains a region (amino acid residues 56-72) that
is highly homologous (88%) to the region of the CDK4 gene (amino
acid residues 9-25) mutated in one melanoma cell line and a familial
melanoma kindred (R24C; Refs. 20 and 26), was performed using
genomic DNA from all 17 of the NB cell lines. This particular CDK4
gene mutation dramatically decreases the ability of CDK4 to bind
pl61NK4abut not other CKIs (e.g., p21 w*i.ciPi.sdii.c.p2o and P27KÃŒP1)
An identical region of CDK4 was also examined by SSCP. The SSCP
analysis of exon 1 from the CDK6 gene indicated that three of the NB
cell lines (NB10, NB12, and NB14) contained possible nucleotide
changes and that none of the CDK4 genes are altered, whereas no
mutations were identified in the remaining (i.e., NB2, NB4, NB5,
NB6, NB8, NB13, NB16, NB17, NB19, NB20, or NB21) cell lines
(data not shown). To definitively identify the alterations in NB10,
NB12, and NB14, genomic PCR products from each NB cell line were
sequenced collectively (Fig. 5A). In addition, the same regions of the
CDK6 gene from NB2, NB6, NB8, NB16, and NB17 were also
examined in the same manner.
The genomic PCR product corresponding to CDK6 from NB14
encoded a partial protein sequence identical to the human CDK6
except for an aspartic acid (D) to tyrosine (Y) replacement at position
32, which was produced as the result of a guanine to thymidine
(G—>T)transversion (Fig. 5ß).Accordingly, this mutation was named
D32Y, and the resulting protein was named CDK6-D32Y. No addi
tional mutations in this region (exon 1, genomic PCR; Fig. 5A) or any
other region of the entire open reading frame (as determined by
RT-PCR; data not shown) were detected in any of the NB cell lines.
Examples of individually isolated CDK6 clones, representative of the
2628
Research.
on September 8, 2019. © 1998 American Association for Cancercancerres.aacrjournals.orgDownloaded from

6. CDK6 MUTATION IN NEUROBLASTOMA
cdk4 IP
CO i- CM 00 ^ IO <O CO !-i--t-T-T-T-i-CMCJ p^ CO
XCÛQÛCÛCÛCÛCÛCÛ CÛOÛCQCÛCÛCÛCÛCÛCÛ *-) co
CCZZZZZZZZZZZZZZZZS O
GST-Rb
cdk6 P
OCMCO-* cor^o) o
i- CM CO ^ in «DOOi-i-1-1- i-i--!- CM
CO CO CO CO CO CÛ CO CÛ CO CO CÛ GuCOCu CÛ
z zzzz zzzzzz zzz z l
Fig. 4. /t, CDK4 and CDK6 protein kinase
activities in the various NB cell lines. B, the effect
of increasing amounts of exogenously added GST-
pl6'NK4'' fusion protein on CDK6 protein kinase
activity from MOLT4 cells (top) and the effect of
5 fig of exogenously added GST-pl6INK4a fusion
protein on CDK6 protein kinase activity from the
various NB cell lines (bottom}. The NB cell lines
used were determined by the results of the CDK6
protein kinase activity assays shown in A. The cell
lines used, as well as the presence ( + ) or absence
(-) of GST-pl6'NK4a protein in the kinase assay,
are shown.
GST-Rb
B
Cdk6 IP
GST-p-16
co
m O 0.10.5 1 5 10 (jig)
GST-Rb
Cdk6 IP
MOLT4 NB2 NB6 NB8 NB10 NB12 NB14 NB16 NB17 C33A
GST-p16 - - +
GST-Rb
pooled population of NB14 CDK6 clones, are also shown (Fig. 5A).
The nucleotide changes in NB10 and NB12 are identical (the result of
an A—»Gtransition of nucleotide 194 of CDK6; GenBank accession
no. X66365) and resulted in a conservative amino acid change, K26R
(data not shown). Consequently, this protein was named CDK6-
K26R. The CDK6-D32Y mutation is directly adjacent (distal) to the
CDK4-R24C mutation that was identified in a melanoma cell line
(Fig. 5D; Ref. 26), whereas the CDK6-K26R mutation is several
residues away (proximal). Both changes occur within the highly
conserved region that is shared by these two CDKs.
Functional Analysis of the CDK6-K26R and CDK6-D32Y Pro
teins. Although no tumor-specific mutation in CDK6 has been re
ported, programmed mutation of arginine residue 31 (identical to the
arginine residue altered in CDK4-R24C) to a cysteine, CDK6-R31C,
has been shown to significantly inhibit pl6INK4a binding to this
protein in vitro (26). Therefore, it was important to determine whether
the K26R and/or D32Y replacements significantly affected the ability
of CDK6 to bind pl6INK4a. Wild-type CDK6, CDK6-K26R, and
CDK6-D32Y proteins were made by IVTT of the appropriate cDNAs
and then incubated with either GST, GST-pl6INK4a, or GST-pl8INK4c
beads in an in vitro binding assay (Fig. 55). As expected, there was no
significant change in the ability of the CDK6-K26R protein to bind to
pjfjiNK4a or pl8INK4c (data not shown), but there was a substantial
decrease (-90%) in the ability of the CDK6-D32Y protein to bind to
the same proteins (Fig. 5B). On the basis of these data, we believe that
the CDK6-K.26R mutation is most likely a polymorphism that does
not effect the ability of the protein to bind pl6INK4a or pl8INK4c.
Finally, to confirm the in vitro binding results as well as to dem
onstrate whether the CDK6-D32Y mutation affected protein kinase
function, we performed in vitro protein kinase assays using either the
wtCDK6 or CDK6-D32Y proteins produced in insect Sf9 cells (Fig.
5Q. This type of analysis was used in a similar study of the effect of
the R24C mutation on CDK4 activity, allowing these investigators to
demonstrate the effect of the CDK4-R24C mutation on protein kinase
activity as well as the ability of exogenously added pl6INK4a to inhibit
such kinase activity (26). Cyclin Dl and wtCDKo, or CDK6-D32Y,
2629
Research.
on September 8, 2019. © 1998 American Association for Cancercancerres.aacrjournals.orgDownloaded from

7. CDK6 MUTATION IN NEUROHL.AS IOMA
L33
Y/D"
R31
G
T
T
C
A
T/G
C
G
C
C
NB14 NB17 NB14
WT
NB14
MUT
G A T C GATC GATC G A T C
(ranging from 50 to 5 jag) were added to both the wtCDKo and
CDK6-D32Y/cyclin Dl kinase complexes to determine whether the
CDK6-D32Y mutation affected the ability of the protein kinase to be
inhibited by this CKI (Fig. 5C). Indeed, although the wtCDKo protein
kinase was inhibited by 50 ng of exogenously added p!6INK4a, the
CDK6-D32Y protein kinase required 1 fig of the same protein for a
similar effect (Fig. 5C, Lanes 2 and 12). Partial inhibition of the
CDK6-D32Y kinase activity was observed when 500 ng of pl6INK4a
was used (Fig. 5C, Lane 11). These results confirm that this mutation
has no effect on protein kinase activity, but that it has a dramatic effect
on the ability of p!6INK4a and pl81NK4c to bind and inhibit CDK6.
B
WT Y32
CO CO (O CO
IVJT O. O. O. O.
, , K K h- H H h-
C/5 C/) C/) £/) CO C/)
WT Y32 O O O O O O
cdk6
12 345 678
p16 O 0-050.1 0.5 1 2.5 5 O 0.050.1 0.5 1 2.5 5 ug
cdk6 WT Y32
-Rb
1 2345678910 11121314
PLSTIRE
1 32 55 326
Fig. 5. Characterization of the CDK6 gene mutation in NB14. A. pooled DNA clones
from NB14 and a nonmutated control. NB17; genomic PCR reactions were sequenced to
identify possible mutations. A T—»Gtransition was found in NB14 that resulted in the
indicated change in the peptide sequence (denoted CDK6-D32Y). Righi panels, these two
reactions are individual clones from the pooled NB14 genomic PCR samples. B. com
parison of the hinding affinity of wild-type CDK6 and CDK6-D32Y for pl6INK4a and
pl8iNKJ« i^ß |«s]methionine-labeled IVTT products of the two clones. CDK6 wild-
type (WT) and the NBI4 CDK6 mutant (K?2), are shown. Pull-down assays were
performed using 5 ^ig of either GST (as a negative control). GST-pl6INK4il. or GST-
plgiNK4c [„.¡„kincubated with the two different IVTT CDK6 products. C protein kinase
activity associated with wild-type ( WT) and mutant (X.Õ2) CDK6 from baculovirus-
expressed proteins. The ability of exogenously added GST-pl6INK4j protein to inhibit
these protein kinase activities is also shown. Increasing amounts of pl6 (0-5 ^ig) were
added to identical cyclin DI/CDK6 protein kinase complexes reconstituted from wild-type
(HT) or mutant (W2) CDK6. Righi, position of the phosphorylated GST-Rb protein
(pRb). D. schematic diagram of CDK6 indicating the location of the mutation in CDK6-
D32Y relative to the previously identified mutation in CDK4 (CDK4-R24C; *) described
previously in melanoma (26).
baculovirus constructs were transfected into Sf9 cells and protein
expression confirmed by Western blotting (data not shown). Equiva
lent amounts of cyclin Dl/wtCDK6 and cyclin D1/CDK6-D32Y
proteins were then used to phosphorylate GST-pRb in vitro, and both
CDK6 proteins appeared to have similar protein kinase activities (Fig.
5C, Lanes I and 8). Increasing amounts of a GST-pl6INK4" protein
DISCUSSION
In this study, we have characterized the cyclin D/CDK/INK/pRb
pathway in a large number of human NB cell lines, many with
amplified MYCN genes and elevated N-myc protein levels. In normal
diploid fibroblasts and growth factor-dependent hematopoietic cells,
elevated levels of c-myc offer a proliferative advantage as long as
growth factors are present (4, 5, 7, 39). However, if growth factors are
removed, these cells undergo apoptosis if elevated c-myc1levels are
maintained. A possible link between mvc and the cyclin D/CDK/INK/
pRb pathway has been proposed, but the data suggest that their
relationship to one another may not be straightforward (39, 40). In
addition, this pathway has been found to be the frequent target of
mutation in many human tumors (Refs. 12, 13, 19, 24, and 25; see Fig.
6). These alterations help these tumor cells disable the normal regu
lation of the cell cycle by growth control mechanisms, promoting
aberrant cell proliferation. Because the nature of the defects affecting
the cyclin D/CDK/INK/pRb pathway in many of these NB cell lines
is currently unknown, we cannot conclude whether elevated N-mvr
levels are relevant.
Several previous studies of this pathway in NB tumors have been
focused on the nature of the p!6INK4u (MTS1) gene, a known tumor
suppressor inactivated by mutation and/or deletion (19, 24). The
results of the majority of these studies demonstrate that the MTSl
gene is not functionally inactivated in NB by either mutation and/or
deletion, although occasional exceptions have been noted (24). In
addition, a study of the integrity of the human /;/#1NK4c gene in a
broad number of tumors, including a number of NBs, indicated that
this gene was rarely altered (31). However, a comprehensive study
addressing the function of the cyclin D/CDK/INK/pRb pathway in NB
has not been available. In agreement with these groups, we have found
that the MTSl and pl8lNK4c genes are normal. We have also found
that their products, pl6INK4u and pl81NK4c. are highly expressed in
>90% of the NB cell lines studied. Such high levels of normal,
functional pl6INK4a and p!81NK4c are not normally associated with
proliferating cells (either nontransformed or transformed), but
pl61NK4a expression has been associated with "terminal" differentia
tion and cellular senescence (41-43).
Accordingly, it has been shown that in the majority of human tumor
cells, a reciprocal relationship between functional pRb and functional
pl61NK4aexists (19, 24). Tumor cells containing high levels of normal
pl61NK4a contain inactive pRb resulting from mutation/deletion of
RBI. rendering these cells insensitive to the effects of normal cell
cycle control exerted at the restriction point prior to entry into S phase.
Conversely, tumor cells identified with homozygous deletion or func
tional mutations in the MTSl gene contain normal RBI genes. Para
doxically, we have found that even in the presence of high levels of
normal pl6INK4a, the RBI gene is normal, and pRb is present in both
its hypo- and hyperphosphorylated forms. pRb is also functional in the
majority of the NB cell lines examined. The D-type cyclin/CDK6
protein kinase activity is markedly elevated in these NB cells, and it
2630
Research.
on September 8, 2019. © 1998 American Association for Cancercancerres.aacrjournals.orgDownloaded from

8. CDK6 MUTATION IN NEUROBLASTOMA
Fig. 6. A schematic representation of observed alter
ations in the cyclin D/CDK/INK/pRb pathway implicated
in various tumors. The diagrammed alterations have been
shown to result in elevated CDK4/6 protein kinase activity
by overcoming the cell cycle block imposed by elevated
pl6INK4u expression. Several possible mechanisms for by
passing normal cyclin D/CDK/INK/pRb control at G,-S
are: elevated cyclin DI levels associated with parathyroid
tumors {/): elevated CDK4 or CDK6 levels associated with
amplification of the respective genes in rhabdomyosarco-
mas. NBs, and gliomas (2); CDK4 or CDK6 gene mutations
(*) that affect the ability of these proteins to interact nor
mally with wild-type pl6INK4ii (3); pl6INK4a (MTSI) gene
mutation and/or deletion (*) associated with many different
tumor types (4); and interaction of pl6INK4;i with a cellular
"factor," such as a v-Tax-relaled protein (c-Tax?). resulting
in its sequestration from active cyclin D/CDK complexes
(5).
Proliferation
t *
Differentiation/Senescence
c-Tax?
Tumor Cells
is possible that this activity contributes to the observed changes in
pRb phosphorylation. Expression of additional CKIs that are also
known to regulate this pathway, p2lw<"'-c>i"™'.r»p*>and p27Klpl,
was invariable and low in all of the NB cell lines, suggesting that these
proteins may not contribute significantly to the regulation of this
pathway in these cells.
Upon further analysis of the individual components that regulate
the cyclin D/CDK/INK/pRb pathway, we identified specific defects in
4 of 17 NB cell lines that are consistent with the results of others,
including a mutation in CDK6 that decreases its protein's sensitivity
to the inhibitory effects of pl61NK4a and pl81NK4c (26). This obser
vation, coupled with the recent report of CDK6 amplification in
gliomas (44), suggests that, like CDK4. CDK6 may be capable of
functioning as a proto-oncogene. A CDK4 gene amplification in
NB21 appears to result in somewhat enhanced steady-state levels of
its protein, but there was no observed effect(s) on CDK4 protein
kinase activity (Figs. 1 and 4). Although it is possible that CDK4 acts
to sequester pl6INK4" and pl81NK4c from complexes containing
CDK6, the presence of both CDK6 and CDK4 in pl6INK4a immune
complexes (Fig. 3) reduces the likelihood of this explanation. In the
remaining 13 NB cell lines, we were not able to identify any defect(s)
in the components of the cyclin D/CDK/INK/pRb pathway that would
explain normal pRb function and elevated CDK6 protein kinase
activity in the presence of such high levels of normal pl61NK4a and
p]8INK4c
How is INK function abrogated in the majority of the NB cell lines
studied? Although the answer to this question is likely to be complex
and possibly multifactorial, we would favor a hypothesis in which an
intrinsic protein factor that binds to, and functionally inactivates,
pl6INK4a is induced in many of these NB cell lines (Fig. 6). Alterna
tively, a mutation in a separate gene product could affect a separate
cell cycle-inhibitory pathway and supersede pl6'NK4l'/pl8INK4c inhib
itory effects. In fact, there is precedence for the first hypothesis. The
oncogenic human T-cell virus HTLV-1 encodes a regulatory protein,
v-Tax, the expression of which can dramatically reduce the amount of
pl6iNK4a that ¡scompiexed witn CDK4/6, allowing cyclin D/CDK4 or
6 protein kinase activation (45). The authors of this study proposed
that such v-Tax:pl6INK4a complexes may contribute to cellular im
mortalization and transformation that is induced by HTLV-1 infec
tion. Amati and colleagues (22) have shown that ectopie expression of
cyclin D1/CDK4 promotes rat fibroblast proliferation in the presence
of elevated pl6INK4a levels by apparently rescuing E2F function via
the pRb pathway. In addition, they show that retrovirally expressed
c-myc can also promote the growth of cells with elevated pl6INK4a,
possibly due to a positive effect of mc on the regulation of cyclin E.
Clearly, whatever the mechanism(s) responsible for the apparent
inactivation of normal pl6INK4a in NB. further analysis of this path
way in these tumor cells will provide significant insight regarding the
regulation of the cell cycle during neuronal cell development, as well
as determine how it contributes to tumorigenesis.
ACKNOWLEDGMENTS
We acknowledge the technical assistance of G. Richmond and J. Crenel, and
M. Valentine for help with the FISH: S. Hiebert for helpful conversations
concerning the EMSA analysis; and D. Quelle for helpful comments regarding
the CDK kinase and pl6INK4a binding experiments. We also acknowledge D.
Beach for providing the GST-pi6INK4a construct: L. Meijer for providing the
human wtCDK6 cDNA: C. Sherr for the human CDK4, p2iw'fl-c'P'-s<'il-c'P20,
and p27Klpl cDNAs; M. Ewen for the GST-pRb construct: and S. Reed for the
human cyclin Dl, D2, and E cDNAs. C. Naeve and the Molecular Resource
Center of St. Jude Children's Research Hospital synthesized the oligonucleo-
tides and assisted with automated DNA sequencing.
REFERENCES
1. Brodeur. G. M.. A/ar, C.. Brother, M., Hiemstra. J. L.. Kaurman. B.. Marshall. H..
Moley. J.. Nakagawara, A., Saylors. R.. Scavard. N., Schneider. S.. Wasson. J..
White. P.. Seeger. R.. Look. A. T., and Castleberry. R. P. Neuroblasloma: effect of
genetic factors on prognosis and treatment. Cancer (Phila.). 70: 1685-1694. 1992.
2. Carón. H. N. Allclic loss of chromosome 1 and additional chromosome 17 material
are hoth unfavourable prognostic markers in neuroblastoma. Med. Pcdiatr. Oncol.,
24: 215-221. 1995.
2631
Research.
on September 8, 2019. © 1998 American Association for Cancercancerres.aacrjournals.orgDownloaded from

9. CDK6 MUTATION IN NEUROBLASTOMA
3. Amati, B., Littlewood, T. D., Evan, G. 1., and Land. H. The c-Myc protein induces
cell cycle progression and apoptosis through dimerization with Max. EMBO J., 72:
5083-5087, 1993.
4. Evan, G. I., Wyllie. A. H., Gilbert, C. S., Littlewood, T. D., Land, H.. Brooks, M.,
Waters. C. M.. Penn, L. Z., and Hancock, D. C. Induction of apoptosis in fibroblasts
by c-myc protein. Cell. 69: 119-128. 1992.
5. Harrington. H. A.. Bennett. M. R.. Fanidi. A., and Evan, G. I. c-Myc-induced
apoptosis in fibroblasts is inhibited by specific cylokines. EMBO J., 13: 3286-3295,
1994.
6. Henriksson, M.. and Luscher. B. Proteins of the Myc network: essential regulators of
cell growth and differentiation. Adv. Cancer Res., 68: 109-182, 1996.
7. Packham, G., Porter, C. W., and Cleveland, J. L. c-Myc induces apoptosis and cell
cycle progression by separable, yet overlapping, pathways. Oncogene, 13: 461-469.
1996.
8. Diecianni. M. B.. Chau. L. S.. Batova. A., Vu, T. Q., and Yu, A. L. The/>/6 and/)/«
lumor suppressor genes in neuroblastoma: implications for drug resistance. Cancer
Lett., 104: 183-192. 1996.
9. Sherr, C. J. G, phase progression: cycling on cue. Cell. 7V: 551-555, 1994.
10. Lukus. J.. Pelersen, B. O., Holm. K.. Bartek, J., and Helin, K. Deregulated expression
of E2F family members induces S-phase entry and overcomes pl6lnk4A-mediated
growth suppression. Mol. Cell. Biol., 16: 1047-1057. 1996.
11. Sherr. C. J., and Roberts. J. M. Inhibitors of mammalian G, cyclin-dependent kinases.
Genes Dev.. 9: 1149-1163, 1995.
12. He, J.. Allen, J. R.. Collins. V. P.. Allalunis-Turner. M. J., Godbout, R.. Day, R. S.,
and James, C. D. CDK4 amplification is an alternative mechanism to ¡riógene
homo/ygous deletion in glioma cell lines. Cancer Res., 54: 5804-5807, 1994.
13. He. J., Olson. J. J.. and James. C. D. Lack of p]6INK4 or retinoblastoma protein (pRb),
or amplification-associated ovcrexpression of cdk4 is observed in distinct subsets of
malignan! glial tumors and cell lines. Cancer Res.. 55: 4833-4836, 1995.
14. Khalib, Z. A., Matsushime, H., Valentine, M.. Shapiro, D. N.. Sherr, C. J., and Look.
A. T. Coamplification of the CDK4 gene with MDM2 and GLI in human sarcomas.
Cancer Res.. 53: 5535-5541, 1993.
15. Koh. J.. Enders. G. H.. Dynlacht. B. D., and Harlow. E. Tumour-derived pió alÃ-eles
encoding proteins defective in cell-cycle inhibition. Nature (Lond.). 375: 506-510,
1995.
16. Medema. R. H.. Herrera. R. E., Lam, F.. and Weinberg. R. A. Growth suppression by
pl6'"k4il requ¡rcsfunctional retinoblastoma protein. Proc. Nati. Acad. Sci. USA, 92:
6289-6293. 1995.
17. Nobori. T.. Miura. K., Wu. D-J., Lois. A.. Takabayashi. K., and Carson, D-A.
IX'letions of the cyclin-dependent kinase-4 inhibitor gene in multiple human cancers.
Nature (Lond.), 368: 753-756, 1994.
18. Okuda. T., Shurtleff, S. A.. Valentine, M. B.. Raimondi. S. C.. Head. D. R.. Behm.
F., Curcio-Brint. A. M., Liu. Q.. Pui, C.. Sherr, C. J.. Beach. D.. Look. A. T., and
Downing. J. R. Frequent deletion of pl6lnk4"/jW752 and pl5Ink4h/Aire/ in pediatrie-
acute lymphoblastic leukemia. Blood. «5:2321-2331. 1995.
19. Parry. D.. Bates. S.. Mann. D. J.. and Peters. G. Lack of cyclin D-Cdk complexes in
Rb-negalive cells correlates with high le
product. EMBO J.. 14: 503-511. 1995.
20. Ranade. K., Hussussian. C. J.. Sikorski. R. S., Varmus. H. E.. Goldstein, A. M.,
Tucker. M. A.. Serrano. M.. Hannon. G. J.. Beach. D.. and Dracopoli. N. C. Mutations
associated with familial melanoma impair pl61NK4 function. Nat. Genet.. 10: 114-
116. 1995.
21. Reed. J. A., Loganzo, F.. Shea. C. R., Walker, G. J., Flores, J. F., Glendening, J. M.,
Bogdany. J. K.. Shiel. M. J.. Haluska. F. G.. Fountain. J. W., and Albino, A. P. Loss
of expression of the/j/6/cyclin-dependent kinase inhibitor 2 tumor suppressor gene in
melanocytic lesions correlates with invasive stage of tumor progression. Cancer Res..
55: 2713-2718, 1995.
22. Alevizopoulos, K., Vlach. J., Hennecke, S.. and Amati. B. Cyclin E and c-Myc
promote cell proliferation in the presence of pl6'"k4" and of hypophosphorylated
retinoblastoma family proteins. EMBO J., 16: 5322-5333. 1997.
23. Berns, K.. Hijmans. E. M.. and Bernards. R. Repression of c-Myc responsive genes
in cycling cells causes G, arrest through reduction of cyclin E/CDK2 kinase activity.
Oncogene, 15: 1347-1356, 1997.
24. Lukas. J.. Aagaard. L., Strauss. M., and Bartek, J. Oncogenic aberrations of pl61NK4/
CDKN2and cyclin Dl cooperate to deregulate G, control. Cancer Res., 55: 4818-
4823. 1995.
25. Lukas. J.. Parry. D., Aagaard. L.. Mann. D. J.. Bartkova. J., Slauss. M.. Peters, G.. and
Bartek. J. Relinoblastoma-protein-dependenl cell-cycle inhibition by the tumour
suppressor pl6. Nature (Lond.), 375: 503-506, 1995.
26. Wolfel. T., Hauer, M.. Schneider. J.. Wolfel, E.. Klehmann-Hieb. E.. DePlaen. E..
Hankeln. T.. Meyer zum Buschenfelde. K. H., and Beach, D. A pl61NK4a-insensitive
CDK4 mutant targeted by cytolytic T lymphocytes in a human melanoma. Science
(Washington DC), 269: 1281-1284, 1995.
27. Lahti. J. M.. Valentine. M.. Xiang. J., Joens. B.. Amann. J.. Grenet. J., Richmond, G..
Look, A. T., and Kidd, V. J. Alterations in the PITSLRE protein kinase gene complex
on chromosome Ip36 in childhood neuroblastoma. Nat. Genet., 7: 370-375, 1994.
28. Scholer. H. R., Hatzopoulos, A. K.. Balling, R., Suzuki, N.. and Grass, P. A family
of octamer-specific proteins present during mouse embryogenesis: evidence for
germline-specific expression of an Oct factor. EMBO J., S: 2543-2550. 1989.
29. Hiebert, S.. Lipp. M., and Nevins. J. R. El A-dependent /r«/i.ï-activationof the human
MYC promoter is mediated by the E2F factor. Proc. Nati. Acad. Sci. USA, «6:
3594-3598, 1989.
30. Hussussian. C. J., Struewing, J. P.. Goldstein, A. M.. Higgins, P. A. T., Ally, D. S.,
Sheahan, M. D., Clark, W. H.. Tucker. M. A., and Dracopoli, N. C. Germline pl6
mutations in familial melanoma. Nat. Genet.. «:15-21. 1994.
31. Zariwala, M.. Liu, E.. and Xiong. Y. Mutational analysis of the pl6 family cyclin-
dependent kinase inhibitors pl5'NK4h and p!8INK4t in tumor-derived cell lines and
primary tumors. Oncogene. 12: 451-455. 1996.
32. Zuo. L., Weger, J.. Yang, Q., Goldstein, A. M., Tucker, M. A., Walker. G. J.,
Hayward, N.. and Dracopoli. N. Germline mutations in the pl6ink4a binding domain
of CDK4 in familial melanoma. Nat. Genet., 12: 97-99, 19%.
33. Kato, J., Matsushime, H.. Hiebert, S. W., Ewen, M. E., and Sherr, C. J. Direct binding
of cyclin D to the retinoblastoma gene product (pRb and pRb phosphorylation by the
cyclin D-dependent kinase CDK4. Genes Dev., 7: 331-342, 1993.
34. Lew, D. J., Dulie, V., and Reed, S. I. Isolation of three novel human cyclins by rescue
of Gl cyclin (Cln) function in yeast. Cell, 66: 1197-1206, 1991.
35. Ohtsubo, M.. Theodoras, A. M., Schumacher, J., Roberts, J. M., and Pagano. M.
Human cyclin E, a nuclear protein essential for the G,-to-S phase transition. Mol.
Cell. Biol. 15: 2612-2624, 1995.
36. Bellinger, C. P.. White. P. S.. Sulman, E. P.. Maris, J. M.. and Brodeur. G. M. No
CDKN2 mutations in neuroblastoma. Cancer Res., 55: 2053-2055. 1995.
37. Liu, Q.. Neuhausen, S., McClure, M.. Frye, C., Weaver-Feldhaus, J., Gruis, N. A..
Eddington. K.. Allalunis-Turner. M. J.. Skolnick. M. H.. Fujimura. F. K.. and Kamb,
A. CDKN2 (MTSI) tumor suppressor gene mutations in human tumor cell lines.
Oncogene. IO: 1061-1067. 1995.
38. Ewen, M. E.. Sluss. H. K.. Sherr, C. J.. Matsushime, H.. Kato. J.. and Livingston.
D. M. Functional interactions of the retinoblastoma protein with mammalian D-type
cyclins. Cell, 73: 487-497. 1993.
39. Roussel, M. F., Theodoras, A. M., Pagano, M., and Sherr, C. J. Rescue of defective
mitogenic signaling by D-type cyclins. Proc. Nati. Acad. Sci. USA, 92: 6837-6841,
1995.
40. Steiner, P.. Philipp. A.. Lukas. J., Godden-Kent. D., Pagano. M., Mittnacht, S.,
Bartek. J.. and Eilers, M. Identification of a Myc-dependent step during the formation
of active G, cyclin-cdk complexes. EMBO J.. 14: 4814-4826. 1995.
41. Alcona, D. A., Xiong. Y., Phelps. D.. Hannon. G., Beach. D.. and Barrett. J. C.
Involvement of the cyclin-dependent kinase inhibitor pl6 (INK4a) in replicative
senescence of normal human fibroblasts. Proc. Nati. Acad. Sci. USA. 93: 13742-
13747, 1996.
42. Lois, A. F., Cooper, L. T., Geng, Y., Nobori, T., and Carson, D. Expression of the pl6
and pi5 cyclin-dependent kinase inhibitors in lymphocyte activation and neuronal
differentiation. Cancer Res.. 55: 4010-4013. 1995.
43. Serrano. M.. Lin, A.. McCurrach, M. E., Beach, D., and Löwe.S. Oncogenic ras
provokes premature cell senescence associated with accumulation of p53 and
pl6iNKJ. Ce|| gg. 593_602, 1997.
44. Costello. J. F.. Plass. C., Arap, W., Chapman, V. M., Held, W. A., Berger, M. S.,
Huang. H. S., and Cavenee. W. K. Cyclin-dependent kinase 6 (CDK6) amplification
in human gliomas identified using two-dimensional separation of genomic DNA.
Cancer Res.. 57: 1250-1254, 1997.
45. Suzuki. T., Kitao. S., Matsushime. H., and Yoshida, M. HTLV-1 Tax protein interacts
with cyclin-dependent kinase inhibitor pl6INK4A and counteracts its inhibitory
activity towards CDK4. EMBO J., 15: 1607-1614, 1996.
46. Serrano, M., Hannon. G. J., and Beach, D. A new regulatory motif in cell-cycle
control causing specific inhibition of cyclin D/CDK4. Nature (Lond.), 366: 704-707.
1993.
2632
Research.
on September 8, 2019. © 1998 American Association for Cancercancerres.aacrjournals.orgDownloaded from

10. 1998;58:2624-2632.Cancer Res
John Easton, Tie Wei, Jill M. Lahti, et al.
Human Neuroblastoma
Kinase/INK4/Retinoblastoma Protein Regulatory Pathway in
Disruption of the Cyclin D/Cyclin-dependent
Updated version
http://cancerres.aacrjournals.org/content/58/12/2624
Access the most recent version of this article at:
E-mail alerts related to this article or journal.Sign up to receive free email-alerts
Subscriptions
Reprints and
.pubs@aacr.orgDepartment at
To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Permissions
Rightslink site.
Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)
.http://cancerres.aacrjournals.org/content/58/12/2624
To request permission to re-use all or part of this article, use this link
Research.
on September 8, 2019. © 1998 American Association for Cancercancerres.aacrjournals.orgDownloaded from