Oncogene (2007) 26, 5115–5123
& 2007 Nature Publishing Group All rights reserved 0950-9232/07 $30.00
ORIGINAL ARTICLE
Epstein–Barr virus promotes genomic instability in Burkitt’s lymphoma
SA Kamranvar1, B Gruhne1, A Szeles and MG Masucci
Department of Cell and Molecular Biology, Karolinska Institutet, Stockholm, Sweden
Epstein–Barr virus (EBV) has been implicated in the
pathogenesis of human malignancies but the mechanisms
of oncogenesis remain largely unknown. Genomic ins-
tability and chromosomal aberrations are hallmarks of
malignant transformation. We report that EBV carriage
promotes genomic instability in Burkitt’s lymphoma (BL).
Cytogenetic analysis of EBV
and EBV þ BL lines and
their sublines derived by EBV conversion or spontaneous
loss of the viral genome revealed a significant increase
in dicentric chromosomes, chromosome fragments and
chromatid gaps in EBV-carrying cells. Expression of
EBV latency I was sufficient for this effect, whereas a
stronger effect was observed in cells expressing latency
III. Telomere analysis by fluorescent in situ hybridiza-
tion revealed an overall increase of telomere size and
prevalence of telomere fusion and double strand-break
fusion in dicentric chromosomes from EBV þ cells.
Phosphorylated H2AX, a reporter of DNA damage and
ongoing repair, was increased in virus-carrying cells in the
absence of exogenous stimuli, whereas efficient activation
of DNA repair was observed in both EBV þ and EBV

cells following treatment with etoposide. These findings
point to induction of telomere dysfunction and DNA
damage as important mechanisms for EBV oncogenesis.
Oncogene (2007) 26, 5115–5123; doi:10.1038/sj.onc.1210324;
published online 26 February 2007
Keywords: EBV; Burkitt’s lymphoma; chromosome
aberrations; genomic instability
Introduction
Epidemiological and molecular evidence implicates
Epstein–Barr virus (EBV) in the pathogenesis of
lymphoid and epithelial cell malignancies (Rickinson
and Kieff, 1996; Dolcetti and Masucci, 2003), but the
mechanisms by which the virus contributes to tumori-
genesis remain largely unknown. EBV infection induces
B-cell immortalization through expression of nine
latency-associated viral proteins, the nuclear antigens
Correspondence: Professor MG Masucci, Department of Cell and
Molecular Biology, Karolinska Institutet, Box 275, Stockholm S-171
77, Sweden.
E-mail: maria.masucci@ki.se
1
These two authors contributed equally to this work.
Received 22 August 2006; revised 7 December 2006; accepted 5 January
2007; published online 26 February 2007
www.nature.com/onc
EBNA1-6 and membrane proteins LMP1, 2A and 2B,
and several viral RNAs of unknown function (Young
and Murray, 2003). The latent viral proteins activate the
cell cycle and inhibit apoptosis but are not sufficient
for full malignancy since recently established EBV
immortalized lymphoblastoid cell lines (LCLs) are
non-tumorigenic, and retain the capacity to undergo
differentiation in lymphoid tissues (Thorley-Lawson,
2001). Furthermore, only some of the growth-transfor-
mation-associated viral proteins are expressed in
EBV-associated malignancies, which led to the identifi-
cation of latency I, where only EBNA1, the untranslated
EBERs and other less well-characterized viral RNAs are
detected, and latency II, where these viral products are
expressed together with LMP1 and LMP2, whereas
EBNA2–6, which characterize the latency III expressed
in LCLs, are downregulated (Dolcetti and Masucci,
2003). EBNA1, the only protein expressed in all latency
types, binds to the origin of latent virus replication
and is required for maintenance of the viral episome
(Yates et al., 1985). Studies addressing the contribution
of EBNA1 to EBV oncogenesis have suggested that it
may act as survival factor, possibly by regulating
p53-induced apoptosis (Hammerschmidt and Sugden,
2004), but have failed to demonstrate conclusively its
direct involvement in malignant transformation.
Genomic instability, defined as the spontaneous
occurrence of high rates of genetic aberrations, is
associated with different types of cancers and appears
to be a major contributor to the neoplastic pheno-
type (Raptis and Bapat, 2006). The changes observed in
tumors include gene mutations, deletions or amplifica-
tions and a variety of karyotype abnormalities collec-
tively defined as chromosomal aberrations. Some
chromosome aberrations, such as translocations, dele-
tions, inversions and duplications, can be transmitted
to progeny cells, whereas unstable chromosomal
aberrations, including dicentric chromosomes, rings,
chromatid gaps, fragments, satellite associations and
double minutes, are not propagated owing to structural
constrains and are generated de novo at each cell cycle.
Both transmissible (Bernheim et al., 1981; Klein, 1986)
and unstable chromosomal aberrations (Stollmann
et al., 1985; Zattara-Cannoni et al., 1996; Chan et al.,
2001) are present in EBV-associated malignancies, but
the contribution of the virus to this phenotype is not
understood. It was suggested that EBV might prime
genomic instability by activating specific recombinases
(Kuhn-Hallek et al., 1995) by integration (Jox et al.,
1997) or by induction of oxidative stress (Gualandi
 EBV promotes genomic instability
SA Kamranvar et al
5116
et al., 2001), whereas latent viral genes such as LMP1
and LMP2 could contribute by modulating telomere
function and DNA repair (Chen et al., 2005; Liu et al.,
2005), or by enhancing the sensitivity to DNA-
damaging agents (O’Nions and Allday, 2003; Liu
et al., 2004). However, there is no conclusive evidence
for the mutagenic potential of the virus.
In this study, we have used a panel of EBV þ and
EBV
Burkitt’s lymphoma (BL) cell lines to investigate
the effect of the virus on genomic instability. Endemic
BL is an aggressive B-cell tumor affecting young
children in endemic malaria regions of Africa and
Papua New Guinea (Dolcetti et al., 2005). The mole-
cular hallmark of BL is the presence of chromosomal
translocations that activate c-myc by juxtaposition to
immunoglobulin enhancer sequences. The importance of
EBV for the pathogenesis of BL is substantiated by
the finding that virtually all endemic BLs are EBV þ .
The tumor biopsies and some derived cell lines express
latency I, whereas less restricted forms of latency are
expressed after prolonged culture in vitro and in in vitro
EBV converted cells. Spontaneous loss of the viral
genome has been reported in rare tumors, providing
additional opportunities for assessing the effect of
the virus in cells with the same genetic background.
We report that EBV carriage is associated with a
significant increase of unstable chromosomal aberra-
tions. The effect was evident in cells expressing latency I,
further enhanced upon progression to latency III and
abrogated in cells that had lost the viral genome,
supporting the involvement of EBV in the induction of
genomic instability.
Results
Enhanced chromosome instability in EBV þ BLs
Non-transmissible chromosomal aberrations including
dicentric chromosomes, chromosome rings, chromo-
some fragments, double minutes, satellite associations
and chromatid gaps (Figure 1A) that arise from
chromosome breakage and fusion, gene amplification,
DNA damage or defects of DNA repair, were scored in
more then 3500 digital images of metaphases from a
panel of 7 EBV
and 20 EBV þ BL lines (Table 1).
More then 100 images captured in at least two different
occasions were analysed for each cell line. Various combi-
nations of chromosomal aberrations were detected in 4.6
and 19.4% of mitosis from EBV
and EBV þ BL lines,
respectively (P ¼ 0.001, Figure 1B). Dicentric chromo-
somes and fragments were observed in a significant

Figure 1 EBV induces chromosomal aberrations in BL cells. (A) Representative metaphase plates and abnormal chromosomes:
(a) normal metaphase plate; (b) metaphase plate with two dicentric chromosomes; (c) dicentric chromosome; (d) chromosome fragment;
(e) chromatid gap; (f) ring chromosome; (g) double minutes; (h) satellite association. (B) Frequency of chromosomal aberrations in
EBV
and EBV þ BL lines; the mean percentages of metaphases showing the indicated chromosomal aberration is shown in the figure.
(C) The originally EBV þ BL lines Oma and Akata were compared with their sublines that have lost the viral genome upon in vitro
culture and the EBV
cell lines BL41, BJAB and Akata(
) were compared with their in vitro EBV converted sublines.

Oncogene
 EBV promotes genomic instability
SA Kamranvar et al
5117
Table 1 Prevalence of chromosomal aberrations in EBV
and EBV+ BL cells
Cell line EBV status EBV proteins EBV latency type % abnormal mitosis

DG75
— — 4
BL28
— — 12
BJAB
— — 6
BJAB-B95.8a Conv EBNA 1–6, LMP1,2 III 11
BJAB-HRIKb Conv EBNA 1, 3–6 I (III)c 14
Ramos
— — 0
Ramos-B95.8a Conv EBNA 1–6, LMP1,2 III 23
Ramos-HRIKb Conv EBNA 1,3–6 I (III)c 3
BL41
— — 6
BL41/B95a Conv EBNA 1–6, LMP1,2 III 68
BL41-E95Aa Conv EBNA 1–6, LMP1,2 III 18
Akata (+) + EBNA 1 I 16
Akata (
)d
— — 2
Akata-Jswt Conv EBNA 1–6 I (III)c 62
Oma cl.6 + EBNA 1 I 10
Oma cl.4d
— — 2
Rael + EBNA 1 I 10
BL72 + EBNA 1–6, LMP1,2 III 8
Namalwa cl.1e + EBNA 1–6, LMP1,2 III 32
Namalwa cl.2e + EBNA 1–6, LMP1,2 III 25
Jijoyef + EBNA 1–6, LMP1,2 III 12
WW2-BLf + EBNA 1–6, LMP1,2 III 8
Raji + EBNA 1–6, LMP1,2 III 17
Mutu cl.216 + EBNA 1 I 5
Mutu cl.148 + EBNA 1 I 7
Mutu cl.99 + EBNA 1–6, LMP1,2 III 32
Mutu cl.176 + EBNA 1–6, LMP1,2 III 7

Abbreviation: EBV, Epstein–Barr virus. aB95.8 EBV strain. bP3HR1 EBV strain. cLMP1 is not detected in these cells. dDerived from the originally
EBV positive Akata(+) and Oma cell lines. eEBV is integrated in chromosome X in Namalwa cl.1 and in chromosome 1 in Namalwa cl.2. fType B
EBV strain.

proportion of the metaphases, whereas chromatid gaps, Ramos (Supplementary Table 2). Similar effects were
double minutes, satellite associations and rings were rare observed in convertants carrying the prototype B95.8 or
events. Chromosome rings and satellite associations the EBNA2 loss mutant, P3HR1-derived, EBV strains.
occurred with the same frequency in EBV
and EBV þ The latter cell lines do not express LMP1, confirming
cells, whereas dicentric chromosomes (P ¼ 0.001), that neither EBNA2 nor LMP1 is essential for the effect.
chromosome fragments (P ¼ 0.001) and chromatid
gaps (P ¼ 0.005) were significantly increased in EBV þ
EBV latency I is sufficient for induction of
cells (Figure 1B, Supplementary Table 2), suggesting a
chromosome instability
prevalence of chromosome breaks and defects in DNA
EBV þ BLs can be classified in EBV latency subgroups
repair. Double minutes were not detected in EBV
but
based on the expression of growth-transformation-
were occasionally found in EBV þ cells. Similar
associated viral proteins (Rowe et al., 1987). A
results were obtained when the analysis was repeated
significant increase of aberrant metaphases from 4.6
on a restricted panel of cell lines after 6 months in
to 9.6% was observed in BL lines expressing latency I
culture (not shown), confirming that the occurrence of
compared to EBV
BLs, whereas a further increase to
chromosomal aberrations is stable over time.
23.6% was detected in cell lines expressing latency III
Comparison of pairs of EBV
and EBV þ cell lines
(Figure 2). The percentage of metaphases with dicentric
with the same genetic background confirmed the
chromosomes increase from 1.1 in the EBV
cells to 3.2
effect of EBV on chromosomal instability (Figure 1C
and 8.7 in EBV þ cells expressing latency I or latency
and Supplementary Table 2). A decrease in aberrant
III, respectively, and similar proportional increases
metaphases from 10% in Oma cl.6 and 16% in
were observed for chromosome fragments and
Akata( þ ) to 2% was observed upon loss of the viral
chromatid gaps. Double minutes were not detected
genome. Dicentric chromosomes, fragments, chromatid
in the EBV
cells, but were present in approximately
gaps and double minutes were significantly decreased,
1% of metaphases from EBV þ cells, irrespective of
whereas a restoration of the defects to levels higher
latency type.
than those observed in the originally EBV þ cells was
observed in Akata-JSWT that was obtained by EBV
conversion of Akata(
) (Figure 1C). Furthermore, Different mechanisms account for the formation
higher frequencies of dicentric chromosomes, fragments, of dicentric chromosomes
chromatid gaps and double minutes were detected in The selective increase of dicentric chromosomes, frag-
EBV converted sublines of BL41, BJAB (Figure 1C) and ments and chromatid gaps in EBV þ cells suggests that
Oncogene
 EBV promotes genomic instability
SA Kamranvar et al
5118
telomere dysfunction, defects in DNA repair or virus-
induced DNA damage, may be involved in this
phenotype. To address this issue, the distribution of
telomeres in isolated and dicentric chromosomes was
investigated by fluorescent in situ hybridization (FISH)
using a specific probe (Figure 3A). Lack or inappro-
priate repair of double-strand breaks (DSB) involving
Figure 2 Expression of the type I EBV latency program is
sufficient for induction of genomic instability. The mean percen-
tages of metaphases showing chromosomal aberrations and the
type of aberrations detected in seven EBV
BL lines, five EBV þ
cell lines expressing latency I and 12 cell lines expressing latency III
(Table 1) are shown in the figure.
telomeres leads to terminal deletions (chromosomes
with less than four telomeres) or ectopic telomere
sequences (chromosomes with more than four telo-
meres) (Figure 3Aa–c). Joining of the sticky ends of two
broken chromosomes will generate dicentric chromo-
somes lacking telomere signals at the fusion site (T0,
Figure 3Ad), whereas unprotected or eroded telomeres
may lead to the formation of dicentric chromosomes
with two detectable telomere signals at the fusion
site owing to telomere–telomere fusion or fusion of
telomeres with the sticky ends of DSB (T2, Figure 3Ae).
In addition, the association of eroded telomeres may
lead to the formation of dicentric chromosomes with
two pairs of telomeric signal at the fusion sites (T4,
Figure 3Af). Four telomere signals were regularly
detected in all chromosomes from mitogen-induced
B-blasts whereas abnormal numbers of telomeres
were present in 37.2% of chromosomes from EBV þ
and 21.8% of chromosomes EBV
BL cells (Figure 3A
and not shown). Analysis of telomere signals at the
fusion sites in at least 20 dicentric chromosomes
from each cell line revealed that 85% of the dicentric
chromosomes in EBV
cells were derived from
telomere association (T4) and only 5% originated
by telomere fusion (T2) and 10% from DSB fusions
(T0) (Figure 3B). In contrast, telomere fusion and DSB
fusions were prevalent in EBV þ cells, where they

Figure 3 Analysis of telomere integrity and telomerase function. Telomere signals were detected by FISH. (A) Representative images
illustrating telomere aberrations (indicated by arrows): (a–c). Isolated chromosomes with one, two, five or no telomere signals.
Dicentric chromosomes derived from: (d) DSB fusion (T0), (e) Telomere fusion (T2), (f) Telomere association (T4). (B) Prevalence of
telomere associations (T4), telomere fusion (T2) and DSB fusions (T0) in dicentric chromosomes from 3 EBV
and 6 EBV þ BL lines.
The number of telomeres signals was scored at the fusion site of 20–50 dicentric chromosomes from each cell line. (C) Expression of
hTERT investigated by Western blot in nuclear (N) and cytoplasmic (c) fractions. One representative experiment out of three
performed with each cell line. (D) The integrated density of telomere signals was analysed in at least 20 metaphase plates from three
independent preparations of normal B-blasts and from the indicated EBV
and EBV þ cell lines using the ImageJ software. The mean
percentages of telomere falling in four intervals of integrated density values was calculated for each cell lines. The results are expressed
as changes relative to the mean values observed in normal B-blasts (% telomeres in the indicated interval in BL cells – % telomeres in
the indicated interval in normal B-blasts).

Oncogene

accounted for 34 and 30% of the dicentric chromo-
somes, respectively.
To assess further the contribution of telomere
dysfunction, the expression of human telomerase reverse
transcriptase (hTERT) was investigated by Western
blot analysis (Figure 3C). In accordance with previous
reports that have failed to demonstrate consistent
EBV-related difference in the expression and activity
of hTERT in BL cells (Mochida et al., 2005), very
similar levels of hTERT were detected in all BL
lines included in our analysis, irrespective of their EBV
carrier status. Furthermore, hTERT was exclusively
detected in nuclear extracts, excluding cytoplasmic
relocalization as a possible cause of functional inactiva-
tion. As telomere dysfunction may occur in the absence
of hTERT defects and shortened or eroded telomeres
may cause the formation of dicentric chromosomes
(Obe et al., 2002), the distribution of telomere sizes was
compared in EBV
and EBV þ cell lines by quantifying
the intensity of the telomere signals using the
ImageJ software (Figure 3D, Supplementary Figure 5).
A different distribution of telomere sizes was revealed by
this analysis. Although the percentage of short telomeres
were increased in EBV
BLs compared with mitogen-
induced B-cell blasts, and long telomeres were corres-
pondingly decreased, the opposite was observed upon
EBV conversion (Figure 3D). Thus, presence of EBV
appears to be associated with a consistent increase of
telomere length.
Chromosome instability correlates with induction
of DNA damage
We then investigated whether EBV may be involved in
the generation of DNA damage. A phosphorylated
variant of histone H2AX, gH2AX-S139, is found in
large chromosomes regions adjacent to DNA breaks,
providing a convenient and sensitive method for
monitoring the formation of DSBs and recruitment of
DNA repair components by immunofluorescence
and Western blot analysis (Lowndes and Toh, 2005).
Representative immunofluorescence staining performed
with a gH2AX-specific antibody is shown in Figure 4a.
A weak nuclear fluorescence, only marginally above the
background detected in normal B-blasts, was detected in
EBV
BLs, whereas a significantly stronger fluorescence
was detected in EBV þ cells (Figure 4a). The effect was
particularly obvious on comparison of Akata(
) with
Akata( þ ) and Akata-JSWT cell lines (Figure 4a). A
significant increase of fluorescence was induced in all cell
lines upon treatment with the DNA-damaging agent
etoposide, confirming that the mechanisms leading to
detection of DSBs and H2AX phosphorylation are
intact. The correlation between EBV carriage and the
presence of spontaneous levels of gH2AX was confirmed
in Western blot analysis as illustrated by representative
blots, where EBV
and converted cell lines were tested
in parallel (Figure 4b). The effect of EBV on the
efficiency of DNA repair was assessed by comparing
the kinetics of gH2AX disappearance after removal of
etoposide (Figure 4c and d). In line with the different
EBV promotes genomic instability
SA Kamranvar et al
5119
levels of expression in untreated cells, a stronger
induction of gH2AX was observed in EBV
compared
to EBV þ cells, 6- and 2-fold increase, respectively
(Figure 4c and not shown). However, a similar kinetics
of gH2AX disappearance was observed upon removal
of etoposide in all cell lines, independently of their
EBV status (Figure 4d). Similar results were obtained
by treatment with the DNA-reducing agent bleomycin
(not shown).
We finally asked whether EBV might be sufficient for
induction of DNA damage in the absence of c-myc
overexpression. To this end, presence of gH2AX was
investigated in the P493-6 cell line that carries a
recombinant EBV with an estrogen-driven EBNA2
and a conditional c-myc gene (Schuhmacher et al.,
1999). In the presence of estrogen and tetracycline, the
growth of P493-6 cells is driven by a latency III gene
expression program (P493-LCL), whereas withdrawal
of the drugs results in switch to latency I-like BL
phenotype with high expression of c-myc (P493-BL).
Significantly higher levels of gH2AX were detected
by immunofluorescence and Western blot analysis
in untreated P493-LCL cells compared with mitogen-
induced B-blasts, whereas a further increase was
observed upon downregulation of EBNA2 and LMP1
and upregulation of c-Myc in P493-BL (Figure 4e). The
expression of gH2AX was increased in all cell types
following etoposide treatment.
Discussion
The results presented in this paper offer new insights
into the mechanisms by which EBV may contribute
to the pathogenesis of BL and, by inference, other EBV-
associated malignancies. Using a simple methodology
based on visual screening of chromosomal abnormalities
in a large number of metaphase plates, we have found
that EBV carriage is associated with the induction of
genomic instability as reflected by a significant increase
of unstable chromosomal aberrations such as dicentric
chromosomes, chromosome fragments and chromatid
gaps (Figure 1). These aberrations are generated in
individual cells and are not transmitted to the progeny
as damaged chromosomes and fragments cannot pro-
perly segregate during mitosis. Random chromosomal
aberrations were previously described in BL cells
(Zech et al., 1976; Zimonjic et al., 2001; Karpova
et al., 2006) and were attributed to the capacity of c-Myc
to induce genomic instability (Potter and Marcu, 1997).
Our data support the predicted role of c-Myc as
unstable chromosomal aberrations were detected in
EBV
BL cells. However, the defects were between 3
and 10 times more frequent in EBV þ cells and a
distinct role of the virus was confirmed by comparison
of paired cell lines that do or do not carry EBV
following in vitro conversion or spontaneous loss of the
viral genome. It is noteworthy that non-transmissible
chromosomal aberrations are also frequent is other EBV-
associated tumors, including nasopharyngeal carcinoma
Oncogene
 EBV promotes genomic instability
SA Kamranvar et al
5120

Figure 4 Detection of DNA damage. (a) The spontaneous and etoposide-induced occurrence of DSB was investigated by gH2AX
immunostaining in EBV
BL lines and EBV þ cell lines expressing latency I or latency III; (b) Western blot analysis of gH2AX
expression in pairs of EBV
and EBV þ BL cells. Extracts from untreated and etoposide-treated B-blasts were included as controls.
One representative experiment out of three. (c) Representative Western blots illustrating the activation of DNA repair following
etoposide treatment as assessed by disappearance of gH2AX. (d) Kinetics of DNA repair in EBV
(BL41, DG75, BJAB, Ramos), EBV
latency I (Mutu cl.148, Rael) and EBV latency III BL lines (Namalwa, Raji, Bjab-B95-8). Cell lines treated with etoposide were chased
in etoposide-free medium. gH2AX expression was detected by Western blot and quantified by densitometry. The mean residual
percentages of gH2AX at the indicated time of chase is shown. One representative experiment out of three. (e) Spontaneous and
etoposide-induced DNA damage detected by gH2AX staining and Western blot analysis in P493 cells expressing LCL-like or BL-like
phenotypes. One representative experiment out of three.

hybrid cell line (Shao et al., 2001; Cao et al., 2002) and
gastric carcinoma (Chan et al., 2002), suggesting that
induction of genomic instability may be a common
mechanism by which EBV contributes to tumor
progression.
The EBV latency program seems to influence the
frequency of chromosomal aberrations in EBV þ BLs.
Thus, abnormal metaphases were more frequent in cells
expressing latency I compared with EBV
cells and
further increased in cells expressing latency III, suggesting
that different viral products may be critical for the
induction and/or amplification of the phenotype (Figure 2).
It is noteworthy that only EBNA1, the untranslated
EBERs RNAs, some poorly characterized transcripts
derived from BamHI A region of the viral genome and
some even less well-characterized micro-RNAs have been
detected in BL cells expressing latency I (Dolcetti and
Masucci, 2003; Pfeffer et al., 2004), suggesting that these
viral products are sufficient for the effect either directly or
through modulation of cellular genes.
Comparison of the type of chromosomal aberrations
detected in EBV
and EBV þ cells suggests two
Oncogene
possible mechanisms by which EBV may promote
genomic instability. The occurrence of dicentric chro-
mosomes, chromosome fragments and chromosomes
with abnormal telomere numbers (Figure 3B) points
to the involvement of DNA breakage. Dicentric
chromosomes with less than four telomere signals at
the fusion site may be generated by non-homologous
DNA repair following DSBs (Obe et al., 2002) or by the
junction or eroded or damaged telomeres (Chan and
Blackburn, 2003), consequent to defects in telomerase
activity or telomere-associated proteins (Smogorzewska
et al., 2002; Urushibara et al., 2004). The virus appears
to interfere with the homeostasis of telomeres as strong
telomere signals were increased in EBV þ compared
with EBV
BLs and normal B-blasts (Figure 3D),
suggesting an increase in telomere length, but the
underlying mechanism and the contribution of this
abnormality to the generation of chromosome breaks
are not clear. Importantly, although LMP2 was
shown to regulate telomerase activity in transfected
epithelial cells (Chen et al., 2005), we could not detect
significant differences in the expression or subcellular

localization of hTERT in EBV
and EBV þ BL cells
(Figure 3C).
In addition to telomere dysfunction, DNA damage
appears to play an important role in the genomic
instability of EBV þ BLs (Figure 4a and b). The
detection of constitutively high levels of gH2AX suggests
that DNA breaks are generated and/or maintained in
EBV þ cells in the absence of exogenous damaging
agents whereas the efficient clearance of gH2AX after
etoposide treatment excludes major defects in the
detection of DNA damage or activation of DNA repair.
Most importantly, low levels of gH2AX were detected in
the P493 LCL in the absence of c-myc overexpression
(Figure 4e), confirming that viral functions are sufficient
for triggering this type of genomic instability. LMP1, the
hallmark of latency III, was shown to induce the
formation of micronuclei in epithelial cells, repress p53-
mediated DNA repair and enhance the sensitivity to
DNA-damaging agents (Liu et al., 2004, 2005). Con-
ceivably, this effect of LMP1 could explain the higher
levels of genomic instability observed in BLs expressing
latency III, but other viral products must cause DNA
damage in cells expressing latency I. EBV infection was
shown to induce micronuclei formation and oxidative
stress in B cells (Gualandi et al., 2001). It remains to be
seen whether this effect may be associated with EBNA1
or other viral products expressed in latency I.
In conclusion, our findings shed a new light on the
oncogenic potential of EBV. It has long been speculated
that EBV infection could generate a pool of pre-malignant
cells that can undergo genetic changes leading to malignant
transformation. Our data suggest that, by inducing
genomic instability, EBV is capable of driving tumor
progression, and, in the absence of a tight control of cells
proliferation, may be sufficient to promote full malignancy.
Material and methods
Cell lines
The cell lines included in this study, their origin, EBV status and
latency type are listed in Table 1 and referenced in Supplemen-
tary Information. The cells were cultured in Roswell Park
Memorial Institute-1640, supplemented with 10–20% fetal calf
serum, 2 mM glutamine, 100 U/ml penicillin and 0.1 mg/ml
streptomycin. EBV gene expression was confirmed by Western
blot using specific polyclonal and monoclonal antibodies (not
shown). Switch between BL-and LCL-like phenotypes was
induced in the P493-6 cell line (Schuhmacher et al., 1999) by
culture in medium supplemented with 0.5 mg/ml tetracycline and
1 mM of b-estradiol, which resulted in upregulation of LMP1
and downregulation of c-Myc below detection levels.
Metaphase plate analysis
Rapidly growing cells were treated with 30 ng/ml colcemide
(KaryoMAX Invitrogen, Paisley, UK) for 90 min to induce
metaphase arrest, washed in hypotonic buffer containing
75 mM KCl, fixed, dropped onto cold glass slides and mounted
in DAPI containing Vectashield (Vector Laboratories,
Burlingame, CA, USA). Digital images were captured using
a LEITZ-BMRB fluorescence microscope (Leica, Wetzlar,
Germany) equipped with a CCD camera (Hamamatsu) and
analysed using the Adobe Photoshop software. Chromosomes
EBV promotes genomic instability
SA Kamranvar et al
5121
containing two centromeres were scored as dicentric, whereas
fusion of short and long arms of the same chromosome was
scored as rings. Isolated chromosome fragments were distin-
guished from double minutes that appear as pairs of tightly
associated fragments. Fusions of fragments with the short
arms of two distinct chromosomes were scored as satellite
association, whereas isolated breaks in one or both chromatids
were scored as chromatid gaps. Statistical significance was
assessed by unpaired, one-tailed Student’s t-test.
Telomere analysis
Telomeres were stained using the Telomere PNA FISH kit,
(DAKO, Glostrup, Denmark). The slides were washed twice
with Tris-buffered saline, treated with proteinase K for 10 min
at room temperature, dehydrated in ethanol and air-dried.
After denaturation at 801C for 5 min, hybridization was
performed at room temperature for 30 min in dark. The slides
were then overlaid with washing solution for 5 min at 651C,
fixed in ethanol and mounted with DAPI-containing Vecta-
shield. The number of telomere signals was counted in digital
pictures and the fluorescence intensity of individual telomeres,
expressed as the product of the telomere area by the average
pixel density, was measured in more then 20 metaphase plates
from each cell line using the ImageJ software (freeware at
http://rsb.info.nih.gov).
Western blotting
Total cell lysates were prepared in lithium dodecyl sulfate
sample buffer, fractionated in precast 4–12% sodium dodecyl
sulfate–polyacrylamide gel electrophoresis gradient gels
(Invitrogen Corp., Carlsbad, CA, USA) and transferred to
polyvinylidene fluoride membranes (Millipore Corporation,
Bedford, USA). The blots were probed with antibodies to
gH2AX-Ser139 (1:1000, Upstate, Lake Placid, USA), hTERT
(1:200, Santa Cruz Biotechnologies Inc), LMP1 (S12, 1:10000),
c-Myc (9E10 1:1000; Santa Cruz Biotechnologies) or b-actin
(AC-15 1:2000, Sigma-Aldrich, St Louis, MO, USA) followed
by the appropriate HRP-conjugates secondary antibody (Zymed,
San Francisco, CA, USA), developed by enhanced chemilu-
minescence and analysed using the Fuji LAS 1000 system
(FujiFilm Medical Systems Inc., Stamford, CT, USA). Signal
intensities were quantified using the Quantity One analysis
system (BioRad Laboratories, Hercules, CA, USA). Nuclear
and cytoplasmic extracts were prepared using the NE-PER
Reagents (Pierce Biotechnologies, Rockford, IL, USA) and the
efficiency of fractionation was confirmed in western blot using
antibodies to a-tubulin (cytoplasmic fraction, GTU-88, Sigma-
Aldrich) and PARP (nuclear fraction C210; Biomol Affinity).
Gamma-H2AX staining
Three  104 cells in 100 ml phosphate buffer saline were
deposited on glass slides by cytospin centrifugation and fixed
in fresh 4% formaldehyde, pH 8 (Merck, Darmstadt,
Germany) for 10 min. The slides were then stained overnight
with anti-gH2AX and developed with goat-anti-rabbit IgG
Alexa Fluor488 (Molecular Probes, Invitrogen Corp.). Digital
images were analysed with Adobe Photoshop. For DNA repair
activation assays the cells were treated with 0.3 mg/ml etopo-
side (Sigma-Aldrich) for 24 h or 50 ng/ml bleomycin (Merck
Biosciences, Darmstadt, Germany) for 5 h and gH2AX
expression was chased in fresh medium.
Abbreviations
BL, Burkitt’s lymphoma; EBV, Epstein–Barr virus; LCL EBV,
immortalized lymphoblastoid cell line; CIN, chromosomal
instability; DSB, double-strand DNA breaks.
Oncogene
 EBV promotes genomic instability
SA Kamranvar et al
5122
Acknowledgements
This study was supported by grants from the Swedish Cancer
Society, the Swedish Medical Research Council and the
References
Bernheim A, Berger R, Lenoir G. (1981). Cytogenetic studies
on African Burkitt’s lymphoma cell lines: t(8;14), t(2;8) and
t(8;22) translocations. Cancer Genet Cytogenet 3: 307–315.
Cao J, Liu Y, Sun H, Cheng G, Pang X, Zhou Z. (2002).
Chromosomal aberrations, DNA strand breaks and gene
mutations in nasopharyngeal cancer patients undergoing
radiation therapy. Mutat Res 504: 85–90.
Chan SW, Blackburn EH. (2003). Telomerase and ATM/Tel1p
protect telomeres from nonhomologous end joining. Mol
Cell 11: 1379–1387.
Chan WY, Chan AB, Liu AY, Chow JH, Ng EK, Chung SS.
(2001). Chromosome 11 copy number gains and Epstein–
Barr virus-associated malignancies. Diagn Mol Pathol 10:
223–227.
Chan WY, Liu Y, Li CY, Ng EK, Chow JH, Li KK et al.
(2002). Recurrent genomic aberrations in gastric carcinomas
associated with H. pylori and Epstein–Barr virus. Diagn Mol
Pathol 11: 127–134.
Chen F, Liu C, Lindvall C, Xu D, Ernberg I. (2005). Epstein–
Barr virus latent membrane 2A (LMP2A) down-regulates
telomerase reverse transcriptase (hTERT) in epithelial cell
lines. Int J Cancer 113: 284–289.
Dolcetti R, Masucci MG. (2003). Epstein–Barr virus: induction
and control of cell transformation. J Cell Physiol 196: 207–218.
Dolcetti R, Serriano D, Masucci MG. (2005). EBV associated
malignancies: epidemiology, pathogenesis and therapeutic
perspective. In: Thebault S (ed). Progress in virus research
Nova Science Publisher Inc: New York, pp 191–226.
Gualandi G, Giselico L, Carloni M, Palitti F, Mosesso P,
Alfonsi AM. (2001). Enhancement of genetic instability in
human B cells by Epstein–Barr virus latent infection.
Mutagenesis 16: 203–208.
Hammerschmidt W, Sugden B. (2004). Epstein–Barr virus
sustains Burkitt’s lymphomas and Hodgkin’s disease. Trends
Mol Med 10: 331–336.
Jox A, Rohen C, Belge G, Bartnitzke S, Pawlita M, Diehl V
et al. (1997). Integration of Epstein–Barr virus in Burkitt’s
lymphoma cells leads to a region of enhanced chromosome
instability. Ann Oncol 8(Suppl 2): 131–135.
Karpova MB, Schoumans J, Blennow E, Ernberg I, Henter JI,
Smirnov AF et al. (2006). Combined spectral karyotyping,
comparative genomic hybridization, and in vitro apoptyping
of a panel of Burkitt’s lymphoma-derived B cell lines reveals
an unexpected complexity of chromosomal aberrations and
a recurrence of specific abnormalities in chemoresistant cell
lines. Int J Oncol 28: 605–617.
Klein G. (1986). Constitutive activation of oncogenes by
chromosomal translocations in B-cell derived tumors. AIDS
Res 2(Suppl 1): S167–176.
Kuhn-Hallek I, Sage DR, Stein L, Groelle H, Fingeroth JD.
(1995). Expression of recombination activating genes
(RAG-1 and RAG-2) in Epstein–Barr virus-bearing B cells.
Blood 85: 1289–1299.
Liu MT, Chang YT, Chen SC, Chuang YC, Chen YR, Lin CS
et al. (2005). Epstein–Barr virus latent membrane protein 1
represses p53-mediated DNA repair and transcriptional
activity. Oncogene 24: 2635–2646.
Liu MT, Chen YR, Chen SC, Hu CY, Lin CS, Chang YT
et al. (2004). Epstein–Barr virus latent membrane protein 1
Oncogene
Karolinska Institute, Stockholm, Sweden. SAK is supported
by a PhD fellowship awarded by the Iranian Government. AS
was supported by personal fellowship awarded by the Cancer
Research Institute/Cancer Foundation, New York, USA.
induces micronucleus formation, represses DNA repair and
enhances sensitivity to DNA-damaging agents in human
epithelial cells. Oncogene 23: 2531–2539.
Lowndes NF, Toh GW. (2005). DNA repair: the importance
of phosphorylating histone H2AX. Curr Biol 15: R99–R102.
Mochida A, Gotoh E, Senpuku H, Harada S, Kitamura R,
Takahashi T et al. (2005). Telomere size and telomerase
activity in Epstein–Barr virus (EBV)-positive and EBV-
negative Burkitt’s lymphoma cell lines. Arch Virol 150:
2139–2150.
O’Nions J, Allday MJ. (2003). Epstein–Barr virus can inhibit
genotoxin-induced G1 arrest downstream of p53 by pre-
venting the inactivation of CDK2. Oncogene 22: 7181–7191.
Obe G, Pfeiffer P, Savage JR, Johannes C, Goedecke W,
Jeppesen P et al. (2002). Chromosomal aberrations: forma-
tion, identification and distribution. Mutat Res 504: 17–36.
Pfeffer S, Zavolan M, Grasser FA, Chien M, Russo JJ, Ju J
et al. (2004). Identification of virus-encoded microRNAs.
Science 304: 734–736.
Potter M, Marcu KB. (1997). The c-myc story: where we’ve
been, where we seem to be going. Curr Top Microbiol
Immunol 224: 1–17.
Raptis S, Bapat B. (2006). Genetic instability in human
tumors. Exs 96: 303–320.
Rickinson AB, Kieff E. (1996). Epstein–Barr virus. In: Fields
BN, Knipe DM, Howley PM (eds). Virology. Lippincott-
Raven: Philadelphia, pp 2397–2446.
Rowe M, Rowe DT, Gregory CD, Young LS, Farrell PJ,
Rupani H et al. (1987). Differences in B cell growth
phenotype reflect novel patterns of Epstein–Barr virus latent
gene expression in Burkitt’s lymphoma cells. EMBO J 6:
2743–2751.
Schuhmacher M, Staege MS, Pajic A, Polack A, Weidle UH,
Bornkamm GW et al. (1999). Control of cell growth by
c-Myc in the absence of cell division. Curr Biol 9: 1255–1258.
Shao JY, Huang XM, Yu XJ, Huang LX, Wu QL, Xia JC
et al. (2001). Loss of heterozygosity and its correlation with
clinical outcome and Epstein–Barr virus infection in
nasopharyngeal carcinoma. Anticancer Res 21: 3021–3029.
Smogorzewska A, Karlseder J, Holtgreve-Grez H, Jauch A, de
Lange T. (2002). DNA ligase IV-dependent NHEJ of
deprotected mammalian telomeres in G1 and G2. Curr Biol
12: 1635–1644.
Stollmann B, Fonatsch C, Havers W. (1985). Persistent
Epstein–Barr virus infection associated with monosomy 7
or chromosome 3 abnormality in childhood myeloprolifera-
tive disorders. Br J Haematol 60: 183–196.
Thorley-Lawson DA. (2001). Epstein–Barr virus: exploiting
the immune system. Nat Rev Immunol 1: 75–82.
Urushibara A, Kodama S, Suzuki K, Desa MB, Suzuki F,
Tsutsui T et al. (2004). Involvement of telomere dysfunction
in the induction of genomic instability by radiation in SCID
mouse cells. Biochem Biophys Res Commun 313: 1037–1043.
Yates JL, Warren N, Sugden B. (1985). Stable replication of
plasmids derived from Epstein–Barr virus in various
mammalian cells. Nature 313: 812–815.
Young LS, Murray PG. (2003). Epstein–Barr virus and
oncogenesis: from latent genes to tumours. Oncogene 22:
5108–5121.

Zattara-Cannoni H, Horshowski N, Figarella-Branger D,
Dufour H, Grisoli F, Vagner-Capodano AM. (1996).
Unusual chromosome abnormalities in primary central
nervous system lymphoma. Leuk Lymphoma 21: 515–517.
Zech L, Haglund U, Nilsson K, Klein G. (1976). Characteristic
chromosomal abnormalities in biopsies and lymphoid-cell
EBV promotes genomic instability
SA Kamranvar et al
5123
lines from patients with Burkitt and non-Burkitt lympho-
mas. Int J Cancer 17: 47–56.
Zimonjic DB, Keck-Waggoner C, Popescu NC. (2001).
Novel genomic imbalances and chromosome translocations
involving c-myc gene in Burkitt’s lymphoma. Leukemia 15:
1582–1588.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc).

Oncogene