EBV+Lymphomas

Shared by: selwinscream

Tags

Epstein-Barr virus, EBV infection, B cells, Epstein Barr Virus, T cells, white blood cells, T cell, tumor cells, Long-Term Success, cell lymphomas

Stats

views:: 8
posted:: 5/2/2011
language:: English
pages:: 10

Public Domain

Document Sample

The new england journal of medicine

review article

mechanisms of disease

Persistence of the Epstein–Barr Virus
and the Origins of Associated Lymphomas
David A. Thorley-Lawson, Ph.D., and Andrew Gross, M.D.

From the Department of Pathology, Tufts
University School of Medicine (D.A.T.-L.),
and the Department of Rheumatology,
Tufts–New England Medical Center (A.G.)
— both in Boston. Address reprint re-
quests to Dr. Thorley-Lawson at the De-
e pstein–barr virus (ebv) is perhaps best known for its ability to
immortalize human B lymphocytes in culture.1 This property makes it a candi-
date for causing human disease, particularly cancer and autoimmune dis-
ease.2,3 Recent work, however, has shown that EBV has evolved strategies that reduce
its potential to become pathogenic.4-6 These new findings have encouraged a reassess-
partment of Pathology, Jaharis Bldg., Tufts
University School of Medicine, 150 Harri- ment of how and when EBV may cause human disease. In this article, we review current
son Ave., Boston, MA 02111, or at david. knowledge of the ways in which EBV establishes and maintains a persistent infection
thorley-lawson@tufts.edu. at the same time that it minimizes its pathogenicity; we also discuss how these charac-
N Engl J Med 2004;350:1328-37. teristics influence the understanding of the role of EBV in lymphomagenesis.
Copyright © 2004 Massachusetts Medical Society.

ebv infection
in vitro and in vivo infection
In vitro, EBV promiscuously infects resting B cells and almost always transforms them
into proliferating blasts. The result is unregulated polyclonal expansion of latently in-
fected lymphoblasts.7,8 The mechanism of this remarkable effect depends on the ex-
pression of nine viral latent proteins that are under the control of a master transcrip-
tion factor, EBV nuclear antigen 2 (EBNA-2).1 The pattern of viral gene expression that
drives this process is called the growth program5 (Table 1). In vitro events are very dif-
ferent from what occurs in the blood of healthy carriers of the virus.3,5 In healthy carri-
ers, the B cells are also latently infected with EBV, but because these cells are all mem-
ory cells9 that are in a resting state,10 they express no viral proteins.11 Cells that express
the growth program are found only in the lymph nodes.12,13 This restriction of the virus
in the blood to resting memory B cells is maintained even in immunosuppressed pa-
tients, in whom the number of virus-infected cells, on average, is 50 times as high as in
immunocompetent patients.14
Primary EBV infection in vivo generally occurs at an early age15 and is usually asymp-
tomatic.16 However, if the infection is acquired during adolescence or later, it can result
in infectious mononucleosis.17 During the early stages of infectious mononucleosis,
extremely large numbers of EBV-infected B cells circulate in the blood, but they are all
resting memory cells (up to 50 percent of such cells may be infected).18 They are not
proliferating blasts and do not enter the growth program.11
Therefore, even under the extreme conditions of infectious mononucleosis or sup-
pression of the immune response, the proliferating-lymphoblastoid stage of viral in-
fection does not occur in the blood, and the infected cells remain in a nonpathogenic
resting state. Why, then, has EBV developed the capacity to transform B cells by means
of the growth program when such cells pose a potential risk to the host? What is the re-
lation of these transformed B cells to the latently infected, resting memory cells in the
peripheral blood? Central to the discussion in this article is the idea that EBV uses a

1328 n engl j med 350;13 www.nejm.org march 25, 2004

The New England Journal of Medicine
Downloaded from www.nejm.org by Julio Girón on November 7, 2010. For personal use only. No other uses without permission.
Copyright © 2004 Massachusetts Medical Society. All rights reserved.
mechanisms of disease

Table 1. Five Transcription Programs Used by EBV to Establish and Maintain Infection.

Type of Infected B Cell* Program Genes Expressed Function of the Program
Naive cell Growth EBNA-1 through EBNA-6, Activates B cell
LMP-1, LMP-2A, and LMP-2B
Germinal-center cell Default EBNA-1, LMP-1, and LMP-2A Differentiates activated
B cell into memory cell
Peripheral-blood memory cell Latency None Allows lifetime persistence
Dividing peripheral-blood EBNA-1 only EBNA-1 Allows viral DNA in latency-
memory cell program cell to divide
Plasma cell Lytic All lytic genes Replicates virus in plasma
cell

* Except where indicated, the types of cell are primarily restricted to the lymphoid tissue of Waldeyer’s ring. EBNA denotes
EBV nuclear antigen, and LMP latent membrane protein.

strategy of transforming latently infected B cells to produce the signals of the germinal center27,28
into proliferating blasts because only in this way that cause latently infected B-cell blasts to form ger-
can the virus convert these cells into long-lived minal centers29 and make the transition into the
memory cells and thereby make the cells nonpatho- memory compartment (Fig. 1).
genic. The latently infected memory cells that have
been produced shut down the expression of viral
the persistence of ebv proteins, enter the latency program11 (Table 1), and
The current model of persistent EBV infection holds circulate primarily between the peripheral blood
that the growth program of the virus activates B cells and Waldeyer’s ring.30 An exception occurs when
to become proliferating blasts so that they can then the latently infected memory cells divide, in which
differentiate into resting memory B cells through the case they express the EBNA-1 protein (Table 1),11
process of the germinal-center reaction (Fig. 1).20,21 thereby allowing viral DNA to replicate.31 In this
It is in the germinal center that an activated naive case, cell division is not driven by the virus, because
B-cell blast that is responding to a foreign antigen none of the growth-promoting latent proteins are
during an immune response undergoes the transi- present, but is instead regulated by the cell as part
tion into a long-lived memory B cell.20,22 The mod- of the normal mechanism of memory B-cell ho-
el stipulates that the difference between the im- meostasis.
mune-activated B-cell blast and the virus-infected Ultimately, the latent virus in memory B cells is
blast that expresses the growth program is that vi- reactivated and replicates at a site that allows it to
ral genes, not antigens, provide all or some of the spread to new hosts. Recent studies have shown
signals required to effect the transition to a memo- that infectious virus is produced when memory cells
ry B cell. in Waldeyer’s ring differentiate into plasma cells
Consistent with this model are the findings that (Fig. 1 and unpublished data). This event closes the
lymphoblasts produced by the growth program re- cycle of viral infection and persistence (Fig. 2) and
semble antigen-activated B cells both in their cell- underscores how extensively EBV makes use of the
surface phenotype23,24 and in their morphologic biology of normal B cells.
features25 and that the only type of EBV-infected In summary, EBV uses its growth program to ac-
B cell that expresses the growth program in the tivate newly infected B cells so that they can differen-
Waldeyer’s ring in healthy carriers is the naive tiate into resting memory B cells. In memory cells
B cell.12,13 The virus is thought to push infected the virus finds a perfect niche. It can persist in them
naive B cells into the memory state by switching the for long periods, because memory B cells rarely die,
cell from the growth program to another pattern of and it is safe in these cells because they express no vi-
transcription called the default program (Table ral proteins that can be detected by the immune sys-
1).5,26 The default program involves the expression tem. Moreover, in cells that are in the resting state,
of only three latent proteins, two of which, latent the virus poses no threat to its host, because the
membrane protein 1 (LMP-1) and LMP-2, are able growth-promoting genes are no longer expressed.

n engl j med 350;13 www.nejm.org march 25, 2004 1329

Normal B-cell differentiation

Antigen T-cell help Survival Cell division
+ antigen signals
T-cell help

Naive B-cell blast GC B cell Post-GC B cell Memory B cell
B cell

Growth
program Default Latency
Virus EBNA-1 only
program program
EBV infection

Figure 1. How EBV May Establish and Maintain Persistent Infection in Memory B Cells.
EBV infects naive B cells that are in the resting state in the lymphoid tissue of Waldeyer’s ring and uses the growth program to activate these
cells to become proliferating blasts. This process parallels the activation of a naive B cell on exposure to an antigen. The antigen-activated
B-cell blast is rescued through entry into the pool of memory B cells when it receives signals from antigen and antigen-specific helper T cells.
The virus switches from the growth program to the default program in order to deliver these rescue signals to the latently infected blast. Then
the memory cells exit the cell cycle and enter the peripheral circulation. For infected cells, entry into the peripheral circulation results in the
shutdown of all protein-encoding genes — known as the latency program. Memory B cells occasionally divide, as part of the homeostatic
mechanism for maintaining stable numbers of cells. When a cell that is carrying the virus undergoes division, it expresses EBV nuclear antigen
1 (EBNA-1) alone to allow the viral genome to divide along with the cell. In response to unknown signals (perhaps bystander T-cell help 19),
memory cells may differentiate into plasma cells and secrete antibody. This differentiation may be related to the mechanisms that sustain life-
time production of antibody. If such a cell contains the virus, it will reactivate viral replication and infectious virus will be produced. GC de-
notes germinal center.

resolution of the infection minimizing the pathogenic effect of ebv
Infection by EBV is controlled by both cellular and A central tenet of the ideas discussed above is that
humoral immune mechanisms (Fig. 2). Antibody EBV uses the strategy of activating latently infected
limits the spread of infectious virus,32 and cytotoxic B cells to become proliferating blasts because this
T cells destroy infected cells that express viral pro- is the only way the virus can convert these cells into
teins.33 The cellular response can be extremely long-lived memory cells. This strategy has impor-
vigorous; in infectious mononucleosis, up to 50 tant implications for the pathogenesis of EBV-
percent of all CD8-positive cells in the blood are associated diseases. On the one hand, activation of
cytotoxic T cells that are directed against cells in the newly infected cell is dangerous to both the
which EBV is replicating.34 The cytotoxic T cells are host and the virus, because it risks the develop-
the main component of the classic lymphocytosis ment of a potentially fatal neoplastic disease that
(atypical lymphocytes) in infectious mononucleo- could limit the period of time in which the virus can
sis.35 It is likely that the major targets for control of spread to new hosts. On the other hand, the virus
EBV by the immune system are memory cells that has the means to ensure that the proliferating lym-
have initiated viral replication. By killing these cells phoblasts are short-lived. In the case of newly in-
and preventing the spread of infectious virus by an- fected naive B cells in Waldeyer’s ring, EBV rapidly
tibody, the immune response reduces the level of pushes the cells out of the cell cycle and into a rest-
infection (Fig. 2). However, the immune system is ing memory state.
unable to eliminate the virus completely, and as a A risk to the host arises if EBV infects a B cell un-
consequence, viral shedding and virus-infected cells der conditions in which the infected cell cannot dif-
persist at a low level, approximately 1 in 10,000 to ferentiate out of the cell cycle or if memory cells are
100,000 memory B cells.30 accidentally triggered into expressing the growth

1330 n engl j med 350;13 www.nejm.org march 25 , 2004

Saliva Tonsil Periphery
Activated blast
Memory cell

Infection Differentiation

Virus

Cell
Immune response
division

Plasma cell

Differentiation

Replication

Figure 2. The Cycle of Persistent EBV Infection.
Every element of the cycle of infection is susceptible to attack by the host’s immune system, with the exception of resting memory B cells, in
which the virus is quiescent. Cytotoxic T cells recognize all other types of infected cells, and antibody neutralizes the virus. Purple lines with
bars denote blocking.

program (as is the case with bystander B cells, Fig. gram in germinal-center cells, and the latency pro-
3). Both cases could lead to uncontrolled growth gram in memory B cells). The relation between
and the development of a tumor. Ordinarily, these such cells has been inferred from parallels with the
possibilities are prevented, because EBV has the biology of normal B cells but has yet to be demon-
apparently paradoxical property of conserving the strated experimentally. Furthermore, the model may
viral targets that the cytotoxic T cells recognize on be oversimplified in several respects. For example,
infected lymphoblasts.6 The conservation of these it implies that viral genes completely replace anti-
antigens by EBV is very different from the continu- gen signaling during the germinal-center reaction.
ous mutations that allow influenza virus and the However, the relative contributions of antigen and
human immunodeficiency virus to avoid recogni- viral genes have not been established. Similarly,
tion by the immune system.36,37 Consequently, cells with a germinal-center phenotype that express
EBV in a proliferating lymphoblast is a sure target the default program have been described, but it is
for the immune response, and conservation of the not known where they reside and whether they ex-
targets guarantees that lymphoblastoid cells that pand to form a true germinal center.
express the growth program but have not differen- An alternative explanation for the persistence of
tiated or cannot differentiate out of the cell cycle EBV infection has been suggested by Kurth et
will be destroyed. al.,38,39 who have proposed that EBV-infected cells
do not participate in the germinal-center reaction
other interpretations but, rather, that EBV directly infects memory B cells.
The ideas discussed here are based on a large body This idea is based on studies in which EBV-infected
of evidence that, in vivo, EBV uses different tran- cells undergoing clonal expansion were identified
scription programs in different types of B cells (the in the germinal centers of tonsils from patients with
growth program in lymphoblasts, the default pro- infectious mononucleosis. The cells expressed

n engl j med 350;13 www.nejm.org march 25, 2004 1331

Figure 3. Putative Checkpoints in the EBV Life Cycle That Might Give Rise to Lymphoma.
EBV normally infects naive B cells in Waldeyer’s ring, which differentiate into memory B cells, exit the cell cycle (thick red arrows), and are there-
fore not pathogenic. Hodgkin’s disease arises from a virus-infected cell that is blocked at the germinal-center stage, which results in constitu-
tive expression of the default program. Burkitt’s lymphoma arises from a germinal-center cell that is entering the memory compartment but is
stuck at the point of proliferation owing to the activated c-myc oncogene. Consequently, the cell expresses EBNA-1 only. In both Hodgkin’s dis-
ease and Burkitt’s lymphoma, the critical event may be a cellular mutation during the immunologic disturbance associated with acute EBV in-
fection. Because the number of infected cells is so high at this point, there is a reasonable possibility that the cell undergoing mutation will have
the virus in it by chance. Any cell other than the naive B cell in Waldeyer’s ring that becomes infected (thin red arrows) and expresses the growth
program will continue to proliferate, because it cannot differentiate out of the cell cycle (thin dashed purple arrows). The rarity of such an event
highlights the extent to which EBV infection is carefully controlled. Normally, bystander B-cell blasts would be destroyed by cytotoxic T cells
(CTL, blue arrow), but if the CTL response is suppressed, the blasts can lead to post-transplantation lymphoproliferative disease (PTLD).

EBNA-2, which is characteristic of the growth pro- naive B cells rapidly differentiate out of the cell cy-
gram. Although this idea is attractive because of its cle by means of the default program to become
simplicity, it provides no mechanism to explain resting memory cells. Owing to the high level of
how the proliferating memory cells exit the cell cy- viremia and the disruption of lymphoid tissue in in-
cle, what the role of the default program may be, fectious mononucleosis, germinal-center or memo-
and why latently infected germinal-center and mem- ry B cells may incidentally become directly infected.
ory cells from the tonsils of healthy carriers do not These bystander infected cells (Fig. 3) will express
express EBNA-2.12 They always express the default the growth program (that is, they will be EBNA-2–
program. positive), not the default program.12 They will then
In fact, the observations of Kurth et al.38,39 are expand rapidly because they cannot differentiate
consistent with and predicted by the model shown out of the cell cycle and will therefore be the domi-
in Figures 1 and 2, which holds that latently infected nant population of infected cells in the tonsils of

1332 n engl j med 350;13 www.nejm.org march 25 , 2004

patients with infectious mononucleosis. Expanding mucosal epithelium (nasopharyngeal and gastric
populations of such infected cells are exactly what carcinoma2,3).
Kurth et al. observed.38,39 Eventually, the cytotoxic
T-cell response destroys these cells, leaving behind ebv and disease
only the small number of infected germinal-center
cells that express the default program — as is seen EBV has been associated with a number of diseases
in healthy carriers of the virus. (Table 2), particularly autoimmune disease and
It has also been suggested that EBV-driven differ- cancer.2,3,45,46 Demonstration of a causative role of
entiation occurs in the extrafollicular regions of the EBV in autoimmune disease has been difficult, how-
tonsil, rather than in the germinal centers. This idea ever, because worldwide, more than 90 percent of
is not inconsistent with the model and was based people are infected with EBV by the time they are
on studies of transgenic mice in which expression adults,15 most of them in early childhood, and EBV
of LMP-1 from a constitutive promoter resulted in persists in them for the rest of their lives. In order
lymphoma and blocked germinal-center develop- to associate EBV with a particular disease, it must be
ment.40 However, LMP-1 is not expressed from a explained why the virus causes disease in only a few
constitutive promoter in the virus,41,42 it is not on- persons, when almost everyone is infected with it.
cogenic in healthy carriers, and in germinal-center Furthermore, because the virus is carried in the
cells it is not expressed alone but, rather, expressed blood by infected memory B cells,9,47 sensitive tests
in the presence of LMP-2,12 which, from experi- will detect it in any inflamed tissue, regardless of the
ments in transgenic mice, is known to cause gut role of the virus in causing the inflammation. The
mucosal B cells to undergo germinal-center devel- model of EBV infection discussed here adds a fur-
opment.29 For these reasons, the phenotype of these ther complication in that EBV is exquisitely sensitive
mice is difficult to interpret, and it can be concluded to changes in the immune system. Changes in the vi-
only that constitutive expression of LMP-1 by itself ral burden or atypical behavior of the virus may be an
blocks germinal-center development. indirect effect of a compromised immune system
that results from the autoimmune disease, rather
ebv and epithelium than evidence that the virus has a role in the disease.
Although the focus of this review is on the infec- The evidence of an association between EBV and
tion of B lymphocytes, it is worth noting that there cancer is stronger than is the case for autoimmune
is increasing evidence to suggest that the epithelium disease, and the ability of the virus to drive cellular
of Waldeyer’s ring also has a role in both primary proliferation identifies it as a potential carcinogen.
infection and viral shedding.43,44 Although the par-
ticipation of epithelium remains to be clearly estab- lymphoma in immunosuppressed patients
lished, it might have major implications for under- Immunosuppressed patients are at risk for B-cell
standing the role of EBV in carcinoma of the lymphoproliferative diseases, such as post-trans-

Table 2. Diseases in Which EBV May Have a Causative Role.

Controversial
Definite Role Strong Association or Discredited Association
Infectious mononucleosis Burkitt’s lymphoma Rheumatoid arthritis
X-linked lymphoproliferative disease Hodgkin’s lymphoma Systemic lupus erythematosus
B-cell lymphoma in patients with Multiple sclerosis
immunosuppressive disease Breast cancer
Nasal NK–T-cell lymphoma* Chronic fatigue syndrome
Nasopharyngeal carcinoma
Leiomyosarcoma in immunosuppressed
patients

* NK denotes natural killer.

n engl j med 350;13 www.nejm.org march 25, 2004 1333

plantation lymphoproliferative disease. These dis- clonal EBV (up to 80 percent in developing coun-
eases are a heterogeneous collection of disorders48 tries).52,53 Although there are no obvious differ-
that usually carry the virus and express the growth ences to indicate that EBV-positive and EBV-nega-
program.49 The obvious explanation for post-trans- tive Hodgkin’s disease are distinct entities, there is
plantation lymphoproliferative disease is that an evidence that infectious mononucleosis is a risk fac-
impaired cytotoxic T-cell response permits unin- tor only for EBV-positive Hodgkin’s disease.54
hibited growth of EBV-infected cells, but the situa- EBV-positive tumors express the viral genes
tion is not so simple. Two events must occur for EBNA-1, LMP-1, and LMP-255-58 of the default tran-
lymphoblastoid cells that express the growth pro- scription program, which is the program that is
gram to survive and evolve into a lymphoma. First, used by latently infected germinal-center B cells.
the EBV-infected cell must be unable to exit the cell The immunoglobulin genes of Reed–Sternberg cells
cycle and become a resting memory B cell. Second, are hypermutated to the same extent as germinal-
the cytotoxic T-cell response must be impeded so center B cells.59 Thus, the immunoglobulin muta-
that the lymphoblasts are not killed. tions and the data on viral gene expression inde-
Post-transplantation lymphoproliferative dis- pendently support the idea that Hodgkin’s disease
ease may be initiated when a type of B cell other arises from an EBV-infected germinal-center B cell.
than a naive B cell in Waldeyer’s ring becomes in- The presence of EBV in approximately 40 per-
fected and expresses the growth program (as in the cent of tumors would seem to rule out a chance as-
case of bystander cells, Fig. 3). These cells cannot sociation of the virus with Hodgkin’s disease, but
exit the growth program, and they continue to pro- this view does not take into account the fact that
liferate owing to the absence of effective T-cell im- the number of EBV-infected cells is extremely high
munity. This mechanism may explain the heteroge- in infectious mononucleosis (up to 50 percent of
neity of the disease.50 Even in immunosuppressed all memory B cells are infected18). Therefore, if the
patients, this event must be rare, because only one or immunologic disruption of infectious mononucle-
two of the millions of EBV-infected cells in the hu- osis, rather than EBV itself, is the risk factor for
man body develop into tumors. That every infected Hodgkin’s disease, there is a high probability that
B cell does not simply proliferate out of control fur- the premalignant germinal-center cell will have
ther attests to the tight regulation of EBV-driven EBV in it by chance.
proliferation in vivo, even in the presence of a crip- A reasonable scenario for the association of
pled immune response. EBV with Hodgkin’s disease is that a germinal-cen-
ter B cell acquires a mutation during infectious
hodgkin’s disease mononucleosis that blocks its differentiation.60 If
The first recognition of an association between that cell happens also to contain EBV, it will be-
EBV and Hodgkin’s disease came from the observa- come a germinal-center cell that constitutively ex-
tion that infectious mononucleosis is a risk factor presses LMP-1 and LMP-2 (Fig. 3 and Table 3). The
for this form of lymphoma.15 Subsequently, Reed– virus may then be carried as a chance passenger, or,
Sternberg cells were found in some cases of infec- more likely, the constitutive expression of LMP-1
tious mononucleosis,51 and approximately 40 per- and LMP-2 will provide growth and survival signals
cent of Hodgkin’s tumors were shown to contain that enhance tumor growth.

burkitt’s lymphoma
Table 3. Putative Infected Cell of Origin and Viral Role for the Three EBV was discovered 40 years ago in tumor cells
EBV-Associated Lymphomas.* from patients who had Burkitt’s lymphoma,61 yet
its contribution to the development of this tumor
Transcription Causative
Type of Lymphoma Putative Cell of Origin Program Role remains enigmatic. The consistent genetic lesion
in Burkitt’s lymphoma is deregulated activation of
Immunoblastic Bystander B cell of any type Growth Likely
the c-myc oncogene owing to reciprocal transloca-
Hodgkin’s Germinal-center cell Default Unsure
tion with an immunoglobulin gene.62-64 Burkitt’s
Burkitt’s Latency-program memory EBNA-1 only Unsure lymphoma has the same pattern of immunoglobu-
B cell stuck in cycle
lin gene hypermutation as germinal-center and
* EBNA denotes EBV nuclear antigen. memory B cells,65 but it has the cellular phenotype
of a germinal-center cell.66

1334 n engl j med 350;13 www.nejm.org march 25 , 2004

The most compelling evidence of the involve- push an EBV-infected cell toward the germinal-
ment of EBV in Burkitt’s lymphoma is the high fre- center phenotype if the genes that promote viral
quency of tumors that carry the virus67 in endemic growth are not expressed.73 This explanation im-
areas (98 percent) and the presence of clonal EBV plies that the virus is present in the tumor cells
in all the tumor cells.68 There is no satisfactory ex- solely by chance, as a passenger. It is difficult, how-
planation of how EBV participates in the patho- ever, to decipher the origin of tumors on the basis
genesis of Burkitt’s lymphoma.69-71 None of the of the phenotype of the end-stage tumor. Given
growth-promoting latent genes are expressed, and that tumorigenesis is a multistep process that oc-
the only latent protein of the virus present is EBNA- curs over long periods of time, it is virtually impos-
1.72 According to current knowledge, an EBNA-1– sible to know how directly the final cellular or viral
only phenotype is present in nontumor cells only phenotype of Burkitt’s lymphoma relates to the
when a latently infected memory cell that expresses original infected precursor.
the latency program divides.11 This mechanism
raises the possibility that Burkitt’s lymphoma aris- conclusions
es if a translocation in the c-myc gene occurs in an
EBV-infected germinal-center cell that is on its way We are beginning to develop a comprehensive un-
to becoming a memory cell. This cell would nor- derstanding of how EBV persists in vivo, and this
mally express the latency program, but owing to knowledge may provide insights into the origin of
the activated c-myc, it is instead stuck in the prolif- EBV-associated diseases. However, the virus has
erating mode and therefore constitutively express- evolved strategies to minimize or eliminate its patho-
es only EBNA-1 (Fig. 3 and Table 3). The mainte- genic potential, in the interest of maintaining in-
nance of the germinal-center phenotype in this fection and the survival of the host in which it per-
presumptive memory-cell tumor can be explained sists. Therefore, causal relationships between the
by the observation that an activated c-myc gene will virus and disease should be interpreted with care.

references
1. Kieff E, Rickinson AB. Epstein-Barr by filtrates of a human leukaemic cell line 15. Henle W, Henle G. Seroepidemiology of
virus and its replication. In: Knipe DM, containing herpes-like virus. Int J Cancer the virus. In: Epstein MA, Achong BG, eds.
Howley PM, eds. Fields virology. 4th ed. 1968;3:857-66. The Epstein-Barr virus. Berlin, Germany:
Vol. 2. Philadelphia: Lippincott Williams 9. Babcock GJ, Decker LL, Volk M, Thor- Springer-Verlag, 1979:61-78.
& Wilkins, 2001:2511-73. ley-Lawson DA. EBV persistence in memory 16. Joncas J, Boucher J, Granger-Julien M,
2. Rickinson AB, Kieff E. Epstein-Barr B cells in vivo. Immunity 1998;9:395-404. Filion C. Epstein-Barr virus infection in the
virus. In: Knipe DM, Howley PM, eds. Fields 10. Miyashita EM, Yang B, Babcock GJ, neonatal period and in childhood. Can Med
virology. 4th ed. Vol. 2. Philadelphia: Lippin- Thorley-Lawson DA. Identification of the Assoc J 1974;110:33-7.
cott Williams & Wilkins, 2001:2575-627. site of Epstein-Barr virus persistence in vivo 17. Niederman JC, McCollum RW, Henle G,
3. Thorley-Lawson DA. Epstein-Barr virus. as a resting B cell. J Virol 1997;71:4882-91. Henle W. Infectious mononucleosis: clinical
In: Austen KF, Frank MM, Atkinson JP, Can- [Erratum, J Virol 1998;72:9419.] manifestations in relation to EB virus anti-
tor H, eds. Sampter’s immunologic dis- 11. Hochberg D, Middeldorp JM, Catalina bodies. JAMA 1968;203:205-9.
eases. 6th ed. Vol. 2. Philadelphia: Williams M, Sullivan JL, Luzuriaga K, Thorley-Law- 18. Hochberg D, Souza T, Catalina M, Sulli-
& Wilkins, 2001:970-85. son DA. Demonstration of the Burkitt’s van JL, Luzuriaga K, Thorley-Lawson DA.
4. Thorley-Lawson DA, Babcock GJ. A lymphoma Epstein-Barr virus phenotype in Acute infection with Epstein–Barr virus tar-
model for persistent infection with Epstein- dividing latently infected memory cells in gets and overwhelms peripheral memory
Barr virus: the stealth virus of human B cells. vivo. Proc Natl Acad Sci U S A 2004;101: B cell compartment with resting, latently
Life Sci 1999;65:1433-53. 239-44. infected cells. J Virol (in press).
5. Thorley-Lawson DA. Epstein-Barr virus: 12. Babcock GJ, Hochberg D, Thorley-Law- 19. Bernasconi NL, Traggiai E, Lanzavec-
exploiting the immune system. Nat Rev son DA. The expression pattern of Epstein- chia A. Maintenance of serological memory
Immunol 2001;1:75-82. Barr virus latent genes in vivo is dependent by polyclonal activation of human memory
6. Khanna R, Slade RW, Poulsen L, et al. upon the differentiation stage of the B cells. Science 2002;298:2199-202.
Evolutionary dynamics of genetic variation infected B cell. Immunity 2000;13:497-506. 20. Liu YJ, Arpin C. Germinal center devel-
in Epstein-Barr virus isolates of diverse geo- 13. Joseph AM, Babcock GJ, Thorley-Law- opment. Immunol Rev 1997;156:111-26.
graphical origins: evidence for immune son DA. Cells expressing the Epstein-Barr 21. MacLennan IC. Germinal centers. Annu
pressure-independent genetic drift. J Virol virus growth program are present in and Rev Immunol 1994;12:117-39.
1997;71:8340-6. restricted to the naive B-cell subset of 22. MacLennan IC, Liu YL, Ling NR. B cell
7. Diehl V, Henle G, Henle W, Kohn G. healthy tonsils. J Virol 2000;74:9964-71. proliferation in follicles, germinal centre
Demonstration of a herpes group virus in 14. Babcock GJ, Decker LL, Freeman RB, formation and the site of neoplastic trans-
cultures of peripheral leukocytes from Thorley-Lawson DA. Epstein-Barr virus- formation in Burkitt’s lymphoma. Curr Top
patients with infectious mononucleosis. infected resting memory B cells, not prolif- Microbiol Immunol 1988;141:138-48.
J Virol 1968;2:663-9. erating lymphoblasts, accumulate in the 23. Thorley-Lawson DA, Nadler LM, Bhan
8. Pope JH, Horne MK, Scott W. Transfor- peripheral blood of immunosuppressed AK, Schooley RT. BLAST-2 [EBVCS], an
mation of foetal human leukocytes in vitro patients. J Exp Med 1999;190:567-76. early cell surface marker of human B cell

n engl j med 350;13 www.nejm.org march 25, 2004 1335

activation, is superinduced by Epstein Barr B cells expanding in germinal centers of Epstein–Barr virus-associated Hodgkin’s
virus. J Immunol 1985;134:3007-12. infectious mononucleosis patients do not disease: epidemiologic characteristics in
24. Thorley-Lawson DA, Schooley RT, Bhan participate in the germinal center reaction. international data. Int J Cancer 1997;70:
AK, Nadler LM. Epstein-Barr virus superin- Proc Natl Acad Sci U S A 2003;100:4730-5. 375-82.
duces a new human B cell differentiation 39. Kurth J, Spieker T, Wustrow J, et al. EBV- 54. Hjalgrim H, Askling J, Rostgaard K, et
antigen (B-LAST 1) expressed on trans- infected B cells in infectious mononucleo- al. Characteristics of Hodgkin’s lymphoma
formed lymphoblasts. Cell 1982;30:415-25. sis: viral strategies for spreading in the B cell after infectious mononucleosis. N Engl J
25. Nilsson K. The nature of lymphoid cell compartment and establishing latency. Med 2003;349:1324-32.
lines and their relationship to the virus. In: Immunity 2000;13:485-95. 55. Oudejans JJ, Dukers DF, Jiwa NM, et al.
Epstein MA, Achong BG, eds. The Epstein- 40. Uchida J, Yasui T, Takaoka-Shichijo Y, et Expression of Epstein-Barr virus encoded
Barr virus. Berlin, Germany: Springer-Verlag, al. Mimicry of CD40 signals by Epstein-Barr nuclear antigen 1 in benign and malignant
1979:225-81. virus LMP1 in B lymphocyte responses. Sci- tissues harbouring EBV. J Clin Pathol 1996;
26. Babcock GJ, Thorley-Lawson DA. Ton- ence 1999;286:300-3. 49:897-902.
sillar memory B cells, latently infected with 41. Fahraeus R, Jansson A, Ricksten A, 56. Deacon EM, Pallesen G, Niedobitek G,
Epstein-Barr virus, express the restricted Sjoblom A, Rymo L. Epstein-Barr virus- et al. Epstein-Barr virus and Hodgkin’s dis-
pattern of latent genes previously found only encoded nuclear antigen 2 activates the viral ease: transcriptional analysis of virus latency
in Epstein-Barr virus-associated tumors. latent membrane protein promoter by mod- in the malignant cells. J Exp Med 1993;177:
Proc Natl Acad Sci U S A 2000;97:12250-5. ulating the activity of a negative regulatory 339-49.
27. Kilger E, Kieser A, Baumann M, Ham- element. Proc Natl Acad Sci U S A 1990;87: 57. Herbst H, Dallenbach F, Hummel M, et
merschmidt W. Epstein-Barr virus-medi- 7390-4. al. Epstein-Barr virus latent membrane pro-
ated B-cell proliferation is dependent upon 42. Wang F, Tsang SF, Kurilla MG, Cohen JI, tein expression in Hodgkin and Reed-Stern-
latent membrane protein 1, which simulates Kieff E. Epstein-Barr virus nuclear antigen 2 berg cells. Proc Natl Acad Sci U S A 1991;88:
an activated CD40 receptor. EMBO J 1998; transactivates latent membrane protein 4766-70.
17:1700-9. LMP1. J Virol 1990;64:3407-16. 58. Niedobitek G, Kremmer E, Herbst H, et
28. Caldwell RG, Wilson JB, Anderson SJ, 43. Tugizov SM, Berline JW, Palefsky JM. al. Immunohistochemical detection of the
Longnecker R. Epstein-Barr virus LMP2A Epstein-Barr virus infection of polarized Epstein-Barr virus-encoded latent mem-
drives B cell development and survival in the tongue and nasopharyngeal epithelial cells. brane protein 2A in Hodgkin’s disease and
absence of normal B cell receptor signals. Nat Med 2003;9:307-14. [Erratum, Nat Med infectious mononucleosis. Blood 1997;90:
Immunity 1998;9:405-11. 2003;9:477.] 1664-72.
29. Casola S, Otipoby KL, Alimzhanov M, et 44. Borza CM, Hutt-Fletcher LM. Alternate 59. Kuppers R, Rajewsky K. The origin of
al. B cell receptor signal strength deter- replication in B cells and epithelial cells Hodgkin and Reed/Sternberg cells in Hodg-
mines B cell fate. Nat Immunol 2004;5:317- switches tropism of Epstein-Barr virus. Nat kin’s disease. Annu Rev Immunol 1998;16:
27. Med 2002;8:594-9. 471-93.
30. Laichalk LL, Hochberg D, Babcock GJ, 45. Cohen JI. Epstein–Barr virus infection. 60. Staudt LM. The molecular and cellular
Freeman RB, Thorley-Lawson DA. The dis- N Engl J Med 2000;343:481-92. origins of Hodgkin’s disease. J Exp Med
persal of mucosal memory B cells: evidence 46. Crawford DH. Biology and disease asso- 2000;191:207-12.
from persistent EBV infection. Immunity ciations of Epstein-Barr virus. Philos Trans 61. Epstein MA, Achong BG, Barr YM. Virus
2002;16:745-54. R Soc Lond B Biol Sci 2001;356:461-73. particles in cultured lymphoblasts from
31. Yates JL, Warren N, Sugden B. Stable 47. Joseph AM, Babcock GJ, Thorley-Law- Burkitt’s lymphoma. Lancet 1964;1:702-3.
replication of plasmids derived from Epstein- son DA. EBV persistence involves strict 62. Klein G. Specific chromosomal translo-
Barr virus in various mammalian cells. selection of latently infected B cells. J Immu- cations and the genesis of B-cell-derived
Nature 1985;313:812-5. nol 2000;165:2975-81. tumors in mice and men. Cell 1983;32:311-5.
32. Pearson G, Dewey F, Klein G, Henle G, 48. Hopwood P, Crawford DH. The role of 63. Leder P. Translocations among antibody
Henle W. Relation between neutralization of EBV in post-transplant malignancies: a genes in human cancer. In: Lenoir GM,
Epstein-Barr virus and antibodies to cell- review. J Clin Pathol 2000;53:248-54. O’Conor GT, Olweny CLM, eds. Burkitt’s
membrane antigens induced by the virus. 49. Thomas JA, Hotchin NA, Allday MJ, et lymphoma: a human cancer model. Lyons,
J Natl Cancer Inst 1970;45:989-95. al. Immunohistology of Epstein-Barr virus- France: International Agency for Research
33. Khanna R, Moss DJ, Burrows SR. Vac- associated antigens in B cell disorders from on Cancer, 1985:341-71. (IARC publica-
cine strategies against Epstein-Barr virus- immunocompromised individuals. Trans- tions no. 60.)
associated diseases: lessons from studies on plantation 1990;49:944-53. 64. Manolov G, Manolova Y. Marker band in
cytotoxic T-cell-mediated immune regula- 50. Timms JM, Bell A, Flavell JR, et al. Tar- one chromosome 14 from Burkitt lympho-
tion. Immunol Rev 1999;170:49-64. get cells of Epstein-Barr-virus (EBV)-posi- mas. Nature 1972;237:33-4.
34. Callan MF, Tan L, Annels N, et al. tive post-transplant lymphoproliferative dis- 65. Klein U, Klein G, Ehlin-Henriksson B,
Direct visualization of antigen-specific CD8+ ease: similarities to EBV-positive Hodgkin’s Rajewsky K, Kuppers R. Burkitt’s lymphoma
T cells during the primary immune response lymphoma. Lancet 2003;361:217-23. is a malignancy of mature B cells expressing
to Epstein-Barr virus in vivo. J Exp Med 1998; 51. Anagnostopoulos I, Hummel M, Kres- somatically mutated V region genes. Mol
187:1395-402. chel C, Stein H. Morphology, immunophe- Med 1995;1:495-505.
35. Wood TA, Frenkel EP. The atypical lym- notype, and distribution of latently and/or 66. Gregory CD, Tursz T, Edwards CF, et al.
phocyte. Am J Med 1967;42:923-36. productively Epstein-Barr virus-infected cells Identification of a subset of normal B cells
36. Phillips RE, Rowland-Jones S, Nixon in acute infectious mononucleosis: implica- with a Burkitt’s lymphoma (BL)-like pheno-
DF, et al. Human immunodeficiency virus tions for the interindividual infection route of type. J Immunol 1987;139:313-8.
genetic variation that can escape cytotoxic Epstein-Barr virus. Blood 1995;85:744-50. 67. de-Thé G. Epstein-Barr virus and Bur-
T cell recognition. Nature 1991;354:453-9. 52. Weiss LM, Movahed LA, Warnke RA, kitt’s lymphoma worldwide: the causal rela-
37. Webster RG, Laver WG, Air GM, Schild Sklar J. Detection of Epstein–Barr viral tionship revisited. In: Lenoir GM, O’Conor
GC. Molecular mechanisms of variation in genomes in Reed–Sternberg cells of Hodg- GT, Olweny CLM, eds. Burkitt’s lymphoma:
influenza viruses. Nature 1982;296:115-21. kin’s disease. N Engl J Med 1989;320:502- a human cancer model. Lyons, France: Inter-
38. Kurth J, Hansmann ML, Rajewsky K, 6. national Agency for Research on Cancer,
Kuppers R. Epstein-Barr virus-infected 53. Glaser SL, Lin RJ, Stewart SL, et al. 1985:165-76. (IARC publications no. 60.)

1336 n engl j med 350;13 www.nejm.org march 25, 2004

68. Gulley ML, Raphael M, Lutz CT, Ross 70. Klein G. In defense of the “old” Burkitt tions in phenotypically distinct clones of a
DW, Raab-Traub N. Epstein-Barr virus inte- lymphoma scenario. In: Klein G, ed. Burkitt’s lymphoma cell line. J Gen Virol
gration in human lymphomas and lym- Advances in viral oncology. Vol. 7. New York: 1990;71:1481-95.
phoid cell lines. Cancer 1992;70:185-91. Raven Press, 1987:207-11. 73. Polack A, Hortnagel K, Pajic A, et al.
69. Lenoir GM, Bornkamm GW. Burkitt’s 71. Kelly G, Bell A, Rickinson A. Epstein- c-myc Activation renders proliferation of
lymphoma, a human cancer model for the Barr virus-associated Burkitt lymphomagen- Epstein-Barr virus (EBV)-transformed cells
study of the multistep development of can- esis selects for downregulation of the nuclear independent of EBV nuclear antigen 2 and
cer: proposal for a new scenario. In: Klein G, antigen EBNA2. Nat Med 2002;8:1098-104. latent membrane protein 1. Proc Natl Acad
ed. Advances in viral oncology. Vol. 7. New 72. Gregory CD, Rowe M, Rickinson AB. Sci U S A 1996;93:10411-6.
York: Raven Press, 1987:173-206. Different Epstein-Barr virus-B cell interac- Copyright © 2004 Massachusetts Medical Society.

n engl j med 350;13 www.nejm.org march 25, 2004 1337