Int. J. Cancer (Pred. Oncol.
): 84, 396–399 (1999)
r 1999 Wiley-Liss, Inc.
ELEVATED SERUM LEVELS OF TRANSFORMING GROWTH FACTOR ␤1
IN EPSTEIN-BARR VIRUS-ASSOCIATED NASOPHARYNGEAL
CARCINOMA PATIENTS
Jingwu XU1, José MENEZES1, U. PRASAD2 and Ali AHMAD1*
1Laboratory of Immunovirology, Department of Microbiology and Immunology,
University of Montreal and Ste-Justine Hospital Research Center, Montreal, Canada
2University of Malaya, Department of Otorhinolaryngology, Pontai Valley, Kuala Lumpur, Malaysia
Nasopharyngeal carcinomas (NPCs) of non-keratinizing
type are strongly associated with Epstein-Barr virus (EBV).
EBV and its gene products induce the synthesis and/or release
of transforming growth factor ␤1 (TGF-␤1) from human cells
and platelets. TGF-␤1 is an immunosuppressive cytokine, and
many tumors are known to secrete it, to counter the host
immune response. To determine the potential role of this
cytokine in the pathogenesis of NPC, 53 serum samples from
patients with EBV-associated NPC and 20 from healthy
donors were analyzed for total and active TGF-␤ content
using ELISA. Serum samples for TGF-␤ content were also
analyzed from NPC patients at different clinical stages of the
tumors. Significantly higher (p F 0.01) levels of active and
total TGF-␤ were found in the sera of NPC patients than in
control sera. The ratio of active:total TGF-␤ was also signifi-
cantly (p F 0.01) increased in the NPC sera. Levels of this
cytokine were also significantly higher (p F 0.05) in the sera
of patients with advanced stages of tumor compared to
patients with earlier stages. Furthermore, higher levels were
seen in patients with relapsing than with remitting tumors;
even higher levels were observed in NPC patients who died of
the tumor. Our data suggest a role of this cytokine in the
pathogenesis of NPC; therefore, it may prove to be a valuable
biomarker molecule for the diagnosis and prognosis of NPC.
Int. J. Cancer (Pred. Oncol.) 84:396–399, 1999.
r 1999 Wiley-Liss, Inc.
Nasopharyngeal carcinoma (NPC), a malignant tumor of the
nasopharyngeal epithelium, is divided into 3 types, i.e., I, II and III,
depending on whether the tumor cells are differentiated, poorly
differentiated or undifferentiated, respectively. Type I NPCs are
rare, occur sporadically all over the world and are not associated
with Epstein-Barr virus (EBV), whereas type II and type III NPCs
(hereafter referred to as NPC) are highly prevalent in Southeast
Asia, southern China, northern and eastern Africa and some North
American indigenous populations. These NPCs are strongly associ-
ated with EBV, though some genetic and environmental factors
have also been documented to play a role in their development (for
review, see Vokes et al., 1997). Without exception, NPC cells
contain EBV and usually express certain EBV antigens and
mRNAs (EBNA-1, LMP1, LMP2A and EBERs). NPC patients
express high titers of anti-EBV antibodies in their sera, of which
IgA is of diagnostic and prognostic value (Vokes et al., 1997).
EBV and its antigens induce transforming growth factor ␤1
(TGF-␤1) from EBV-infected cells (Caryol and Flemington, 1995),
and we have demonstrated the release of this cytokine from human
platelets after their interaction with EBV (Ahmad and Menezes,
1997). TGF-␤ is produced by almost every cell in the body, and
many tumor cells actively secrete this cytokine. In fact, it has been
suggested as a useful biomarker for certain tumors, e.g., breast
cancers, prostate cancer and hepatocellular carcinoma (Ivanovic et
al., 1995; Shirai et al., 1994). However, its concentrations in the
sera of NPC patients have not been reported. While studying
immune responses and cytokine profiles in different EBV-
associated disease conditions, we found markedly elevated levels
of TGF-␤1 in the sera of NPC patients compared to healthy
EBV-seropositive controls. These findings are reported here.
Publication of the International Union Against Cancer
Publication de l’Union Internationale Contre le Cancer
MATERIAL AND METHODS
Subjects
Fifty-three serum samples were collected from patients having
EBV-associated, undifferentiated or poorly differentiated NPC
from referral centers in Malaysia, where NPC is highly prevalent.
All of these patients had high titers of EBV-specific IgA in their
sera. Sera were analyzed for active and total TGF-␤ using a
commercial ELISA kit (TGF-␤1 Emax Immuno Assay System;
Promega, Madison, WI), following the manufacturer’s recommen-
dations. In this assay, biologically active TGF-␤1 is determined in
an antibody sandwich format. The first antibody used to coat the
plates (capture antibody) is a monoclonal antibody that binds
specifically to active TGF-␤1. For detection of the total biologi-
cally processed form of this cytokine, the same kit was used but the
samples are first acid-treated and neutralized (which renders all
latent TGF-␤1 into its active form). Each determination was
repeated until absorbance values within the range of the standard
curve were obtained. The patients were further classified into
different tumor stages based on tumor size (T), lymph node
involvement (N) and presence of hematogenous metastasis (M)
following the modified Ho stage classification (Teo et al., 1991).
We also monitored TGF-␤ levels in the sera of 15 NPC patients, of
whom 5 had remissions, 5 had relapsing tumors and 5 died of NPC.
All of these patients had undergone radiation and chemotherapy.
Three samples were collected from each patient at 6-month
intervals, except in cases where samples could not be taken for the
third time due to death of the patient.
Statistical analysis
Mean values between the groups were compared using Student’s
unpaired 2-tailed t-test with Instat-2 software (Graph Pad Software,
San Diego, CA).
RESULTS
Individual concentrations of total TGF-␤ and of its active form
found in the sera of healthy individuals and NPC patients are
depicted in Figure 1. Their average values along with the statistical
significance of the differences between means are given in Table I.
It is clear from these data that NPC patients have significantly
higher (p ⬍ 0.01) levels of both active and total TGF-␤ in their
sera. On average, NPC sera had 6-fold more active TGF-␤ than
normal EBV-seropositive control sera, while the level of total
TGF-␤ was approximately 3-fold higher than in control samples.
The mean ratio of active:total TGF-␤ was also significantly higher
in NPC patients (Table I).
Grant sponsors: Medical Research Council of Canada; J.-Louis Lévesque
Foundation.
*Correspondence to: Laboratory of Immunovirology, Ste-Justine Hospi-
tal, 3175 Côte Ste-Catherine, Montreal, Québec, H3T 1C5 Canada. Fax:
⫹1–514–345–4801. E-mail: ahmada@justine.umontreal.ca
Received 14 December 1998; Revised 12 March 1999

ELEVATED SERUM LEVELS OF TGF-␤ IN NPC PATIENTS
Average values for patients with different tumor stages are given
in Table II. Significantly higher (p ⱕ 0.05) levels are found in
advanced compared with earlier tumor stages. Average values for
active and total TGF-␤ for patients with different clinical condi-
tions of the tumor are given in Table III. Mean levels of TGF-␤
correlated with the clinical condition of the patients and differed
significantly between the 3 groups; higher levels were seen in
patients with relapses or who later died of the tumor.
DISCUSSION
The TGF-␤ family is a group of closely related polypeptides
produced by a variety of cells in the body that can display a wide
array of activities depending on the cell type and culture conditions
(reviewed by Stravnezer, 1995). Receptors for TGF-␤ have been
FIGURE 1 – Fifty-three serum samples taken from EBV-associated
NPC patients and 20 from age-matched healthy EBV-seropositive
individuals were assayed for quantitation of total and active TGF-␤1
contents using a commercial ELISA kit (Promega). The figure shows
the individual values of total and active TGF-␤1 in healthy controls (A)
and NPC patients (B). Horizontal bars denote average values of the
group.
TABLE I – COMPARISON OF THE SERUM TGF-␤1 LEVELS BETWEEN SERA OF NPC PATIENTS AND HEALTHY,
EBV-SEROPOSITIVE (CONTROL) INDIVIDUALS
Number of Active
samples (mean ⫾ SE)
NPC patients 53 35.91 ⫾ 2.29
⬍0.0001
Controls 20 6.69 ⫾ 0.61
Serum samples collected from NPC patients and control subjects were analyzed for total and active
TGF-␤1 contents using a kit (Promega). Table shows means ⫾ SE of the 2 groups and the ratio of the
active:total TGF-␤ along with the statistical significance of the differences between means.
10970215, 1999, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/(SICI)1097-0215(19990820)84:4<396::AID-IJC11>3.0.CO;2-# by Nat Prov Indonesia, Wiley Online Library on [13/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
397
found on almost every cell line tested so far, which enables this
cytokine to exert its effects on almost every body tissue (reviewed
by Massagué, 1992). It is the most immunosuppressive substance
found to date in the human body, blunting the cellular (T and NK
cells) immune response and inducing the B cells, alone or in
conjunction with other cytokines (e.g., IL-10), to produce IgA
(Stravnezer, 1995). It inhibits the growth and proliferation of T, B
and epithelial cells and plays a role in tumor development by acting
as an angiogenic factor as well as promoting metastases (Stravn-
ezer, 1995). Since many tumor cells have been reported to produce
TGF-␤, it is not surprising that elevated concentrations of this
cytokine are found in the blood of patients with different tumors,
e.g., breast cancer, prostate carcinoma and hepatocellular carci-
noma (Ivanovic et al., 1995; Shirai et al., 1994; Kong et al., 1995).
However, in the case of NPC, elevated TGF-␤ levels may occur
due both to its secretion from tumor cells and to EBV infection,
which is known to enhance TGF-␤ secretion (Cayrol and Fleming-
ton, 1995; Ahmad and Menezes, 1997). Keeping in view other
tumor cells, it is highly likely that NPC cells may themselves
secrete TGF-␤. High TGF-␤ levels may play a special and unique
role in the pathogenesis of NPC.
TGF-␤ is secreted in a latent form, which is activated by a poorly
understood mechanism before interacting with ubiquitously occur-
ring TGF-␤ receptors on cells. Under physiological conditions, the
concentration of the active form is usually less than one-third of
total TGF-␤, as observed here in the control sera. However, the
active form of TGF-␤ is more than half of the total value in the
NPC sera. Thus, in NPC patients, TGF-␤ not only is secreted more
but also appears to be more rapidly activated. A viral protein,
influenza virus neuraminidase, has been reported to activate latent
TGF-␤ (Cherry and Hinshaw, 1996). Further work is needed to
determine if any of the EBV gene products, which are expressed at
higher rates in these patients, has this property.
EBV encodes a viral homolog of human IL-10, BCRF1 orf
(Moore et al., 1993); interestingly, TGF-␤ induces B cells to switch
to IgA production in conjunction with IL-10. In this regard, NPC
patients exhibit high levels of EBV-specific IgA in their sera (Vokes
et al., 1997). This IgA is of diagnostic value in these patients, and
its rise represents a useful marker for detecting early occult cases of
NPC.
TGF-␤ also causes disruption of viral latency and stimulates
viral replication in Burkitt’s lymphoma cells (Renzo et al., 1994).
Thus, high levels of TGF-␤ may, at least in part, be responsible for
high viral replication seen in NPC patients. Localized replication of
EBV in the nasopharyngeal epithelium of NPC patients would not
only lead to increased shedding of virus in the saliva but may also
result in a local (mucosal) immune response that produces anti-
EBV IgA. IgA protects mucosal surfaces from invading pathogens
and clears immune complexes from the circulation. However, in
the context of EBV infection and NPC, EBV-specific IgA may
contribute toward pathogenesis. EBV-specific IgA mediates infec-
tion of epithelial cell lines that produce secretory component (SC)
and, thus, can transcytose IgA (Sixbey and Yao, 1992). There is
strong circumstantial evidence that this route of epithelial cell
infection may be important in vivo since the epithelial cells that
produce SC are localized in the fossa of Rosenmuller, where NPC
develops and localized EBV infections occur (Nomori et al.,
Total Ratio
p value (mean ⫾ SE) p value (active:total)
64.86 ⫾ 6.61 0.554
0.0002
21.64 ⫾ 0.76 0.309
 10970215, 1999, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/(SICI)1097-0215(19990820)84:4<396::AID-IJC11>3.0.CO;2-# by Nat Prov Indonesia, Wiley Online Library on [13/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
398 XU ET AL.

TABLE II – COMPARISON OF SERUM TGF-␤1 LEVELS BETWEEN PATIENTS WITH DIFFERENT TUMOR STAGES

Active Total
Status Number p value p value
(mean ⫾ SE) (mean ⫾ SE)

Tumor stage
T1 27 28.33 ⫾ 1.96 53.91 ⫾ 3.19
T2 5 35.66 ⫾ 4.39 59.37 ⫾ 7.03
T3 9 40.34 ⫾ 2.73 74.66 ⫾ 5.74
T4 12 42.22 ⫾ 5.38 81.05 ⫾ 7.64
T1–T2 32 29.48 ⫾ 1.83 54.76 ⫾ 2.89
T3–T4 21 41.41 ⫾ 3.23 0.00055 78.31 ⫾ 4.95 ⬍0.0001
Lymph node involvement
N0 4 25.20 ⫾ 1.92 41.54 ⫾ 2.18
N1 6 22.21 ⫾ 1.12 50.72 ⫾ 1.02
N2 35 36.49 ⫾ 2.59 64.76 ⫾ 4.12
N3 8 40.68 ⫾ 2.72 79.31 ⫾ 2.55
N0–N1 10 23.41 ⫾ 1.07 46.76 ⫾ 2.28
N2–N3 43 37.28 ⫾ 2.18 0.0019 64.46 ⫾ 3.52 0.0105
Metastasis
M0 36 30.35 ⫾ 1.79 58.32 ⫾ 3.01
M1 17 42.39 ⫾ 4.29 0.0017 74.66 ⫾ 5.86 0.0041
Patients were classified into different T, N and M stages based on a modified Ho stage classification (Teo
et al., 1991). Mean TGF-␤ levels between early and advanced stages were statistically compared, as
described in ‘‘Material and Methods’’. Statistically significant differences between early and advanced
stages for both total and active TGF-␤ are evident.

TABLE III – COMPARISON OF SERUM TGF-␤1 LEVELS BETWEEN PATIENTS functional in colorectal carcinomas, and pancreatic tumor cells
WITH RELAPSES AND REMISSIONS OF TUMOR AND DEATHS
have mutations in DPC4 (SMAD-4), an important component of
Patients with Number of Active Total the TGF-␤ signaling pathway (Markowitz et al., 1995; Kadin et al.,
samples (mean ⫾ SE) (mean ⫾ SE)
1994; Zhang et al., 1997). It is not known which strategy NPC
a: Remissions 15 27.69 ⫾ 2.69 45.82 ⫾ 4.11 tumors have developed to escape high serum levels of TGF-␤.
b: Relapses 15 35.98 ⫾ 2.21 72.64 ⫾ 4.76 Since NPC tumor cells are EBV genome–positive and frequently
c: Deaths 10 50.67 ⫾ 5.22 90.26 ⫾ 6.24 express LMP1, which has been shown to confer refractoriness to
p values the effects of TGF-␤ in transfected cells (Blomhoff et al., 1987;
a vs. b 0.024 0.002 Arvanitakis et al., 1995), they may not require additional mecha-
a vs. c 0.0003 ⬍0.0001 nisms for evasion from TGF-␤. Clearly, further work is needed to
b vs. c 0.0075 0.0329
answer these questions.
Serum samples were collected from 15 patients, of whom 5 had We have described elevated serum levels of TGF-␤ in patients
relapses, 5 had remissions of the tumor and 5 died. Samples were with EBV-associated NPC and their relationship to different tumor
collected 3 times after every 6 months except in the 5 patients who died stages as well as to clinical conditions of the patients. These
before the samples could be collected for the third time. The mean ⫾
SE for each group, for total and active TGF-␤1, is given along with findings, in addition to contributing to the current understanding of
statistical significance of the differences between the mean values of the pathogenesis of NPC, may have implications for the treatment
the 3 groups. of this cancer. These tumors are usually treated by surgery and
irradiation; however, relapses and complications are common.
Strategies to neutralize the excessive amounts of circulating TGF-␤
1985). Thus, elevated serum TGF-␤ with or without EBV-specific in these patients may slow tumor progression and relapses, as has
IgA may contribute toward increased EBV replication, TGF-␤ also been proposed for other disease conditions (Border and Noble,
production and NPC development. 1995).
TGF-␤ inhibits the growth of epithelial and several other cell
types (Stravnezer, 1995). Elevated TGF-␤ may thus inhibit the ACKNOWLEDGEMENTS
proliferation of tumor cells. However, several tumors have adopted
strategies to escape from the growth-inhibitory effects of this We are grateful to Dr. G. Aronheim for comments on the
cytokine; e.g., TGF-␤ type II receptors are mutated and non- manuscript and to Ms. M. Patenaude for secretarial assistance.

REFERENCES
AHMAD, A. and MENEZES, J., Binding of the Epstein-Barr virus to human
platelets causes the release of transforming growth factor-␤1. J. Immunol.,
159, 3984–3988 (1997).
ARVANITAKIS, L., YASEEN, N. and SHARMA, S., Latent membrane protein-1
induces cyclin D2 expression, pRb hyperphosphorylation, and loss of
TGF-␤-1-mediated growth inhibition in EBV-positive B cells. J. Immunol.,
155, 1047–1056 (1995).
BLOMHOFF, H.K., SMELAND, E., MUSTAFA, A.S., GODAL, T. and OHLSSON,
R., Epstein-Barr virus mediates a switch in responsiveness to transforming
growth factor, type beta, in cells of the B cell lineage. Europ. J. Immunol.,
17, 299–301 (1987).
BORDER, W.A. and NOBLE, N.A., Targeting TGF-␤ for treatment of disease.
Nature (Med.), 1, 1000–1001 (1995).
CAYROL, C. and FLEMINGTON, E.K., Identification of cellular target genes of
the Epstein-Barr virus transactivator Zta: activation of transforming growth
factor beta igh3 (TGF-␤ igh3) and TGF-␤1. J. Virol., 69, 4206–4212
(1995).
CHERRY, S.S. and HINSHAW, V.S., Influenza virus neuraminidase activates
latent transforming growth factor ␤. J. Virol., 70, 8624–8629 (1996).
IVANOVIC, V., MELMAN, A., DAVIS-JOSEPH, B., VALCIC, M. and GELIEBTER, J.,
Elevated plasma levels of TGF-␤1 in patients with invasive prostate cancer.
Nature (Med.), 1, 282–284 (1995).
KADIN, M.F., CAVAILLÉ-COLL, M.W., GERTZ, R., MASSAGUÉ, J., CHEIFETZ,
S. and GEORGE, D., Loss of receptors for transforming growth factor ␤ in
human T-cell malignancies. Proc. nat. Acad. Sci. (Wash.), 91, 6002–6006
(1994).
KONG, L.M., ANSCHER, M.S., MURASET, T., ABBOTT, D.A., IGLEHART, J.D.
and JIRTLE, R.L., Elevated plasma transforming growth factor-␤1 levels in

ELEVATED SERUM LEVELS OF TGF-␤ IN NPC PATIENTS
breast cancer patients decrease after surgical removal of the tumor. Ann.
Surg., 222, 155–162 (1995).
MARKOWITZ, S., WANG, J. and MYEROFF, L., Inactivation of the type II
TGF-␤ receptor in colon cancer cells with microsatellite instability.
Science, 268, 1336–1338 (1995).
MASSAGUÉ, J., Receptors for the TGF-␤ family. Cell, 69, 1067–1070
(1992).
MOORE, K.W., O’GARA, A., DEWAAL-MALEFYT, R., VIERRA, P. and MOSS-
MANN, T.R., Interleukin 10. Ann. Rev. Immunol., 11, 165–190 (1993).
NOMORI, H., KAMEYA, T., SHIMOSATO, Y., SAITO, H., EBIHARA, S. and ONO,
I., Nasopharyngeal carcinoma: immunohistochemical study for keratin,
secretory component and leukocyte common antigen. Jpn. J. clin. Oncol.,
15, 95–105 (1985).
RENZO, L.D., ALTIOK, A., KLEIN, G. and KLEIN, E., Endogenous TGF-␤
contributes to the induction of the EBV lytic cycle in Burkitt lymphoma cell
lines. Int. J. Cancer, 57, 914–919 (1994).
10970215, 1999, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/(SICI)1097-0215(19990820)84:4<396::AID-IJC11>3.0.CO;2-# by Nat Prov Indonesia, Wiley Online Library on [13/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
399
SHIRAI, Y., KAWATA, S., TAMURA, S., ITO, N., TSUSHIMA, H. and TAKAISHI,
K., Plasma transforming growth factor-␤1 in patients with hepatocellular
carcinoma: comparison with chronic liver diseases. Cancer, 73, 2275–2279
(1994).
SIXBEY, J.W. and YAO, Q.Y., Immunoglobulin A-induced shift of Epstein-
Barr virus tissue tropism. Science, 255, 1578–1580 (1992).
STRAVNEZER, J., Regulation of antibody production and class switching by
TGF-␤. J. Immunol., 155, 1647–1651 (1995).
TEO, P.M.L., TSAO, S.Y., HO, J.H.C. and YU, P., A proposed modification of
the Ho stage-classification for nasopharyngeal carcinoma. Radiol. Oncol.,
21, 11–23 (1991).
VOKES, E.E., LEIBOWITZ, D.N. and WEICHSELBAUM, R.R., Nasopharyngeal
carcinoma. Lancet, 350, 1087–1091 (1997).
ZHANG, Y., MUSCI, T. and DERYNCK, R., The tumor suppressor Smad4/
DPC4 as a central mediator of Smad function. Curr. Biol., 7, 270–276
(1997).