Hematopathology / EBER AND LMP1 INTERPRETATION GUIDELINES
Guidelines for Interpreting EBER In Situ Hybridization
and LMP1 Immunohistochemical Tests for Detecting
Epstein-Barr Virus in Hodgkin Lymphoma
Margaret L. Gulley, MD,1 Sally L. Glaser, PhD,2 Fiona E. Craig, MD,3 Michael Borowitz, MD, PhD,4
Risa B. Mann, MD,4 Sarah J. Shema, MS,2 and Richard F. Ambinder, MD, PhD4
Key Words: Epstein-Barr virus; Hodgkin lymphoma; EBER; In situ hybridization; LMP1; Immunohistochemistry
Abstract
Histochemical stains demonstrate Epstein-Barr
virus (EBV) in approximately 40% of all Hodgkin
lymphomas, suggesting a role in tumorigenesis and the
potential for EBV-targeted therapy. As research
progresses, it is important to define criteria for
interpreting histochemical stains. Four hemato-
pathologists independently interpreted EBV-encoded
RNA (EBER) and latent membrane protein 1 (LMP1)
histochemical stains from 40 cases of Hodgkin
lymphoma and then reviewed the stains as a group to
resolve discrepancies and to develop interpretation
guidelines. To call a Hodgkin case EBV-related, the
EBER and/or LMP1 signal must be unequivocally
present in Reed-Sternberg/Hodgkin (RS/H) cells. The
cytologic features and distribution of stained cells
should be matched with those on the corresponding
H&E-stained slide to help interpret whether the EBER
or LMP1 signal is in malignant or reactive cells. The
EBER signal is localized to the nucleus, whereas LMP1
is in the cytoplasm and surface membrane. In some
cases, only a fraction of RS/H cells express these factors
for technical or biologic reasons. Before calling a case
EBER-negative, it is essential to show that tumor cell
RNA is preserved and available for hybridization.
LMP1 staining, although usually strong among all
tumor cells in a given case, may alternatively be focal
and weak, contributing to false-negative interpretation.
EBER and LMP1 assays in combination are more
effective than either assay alone for identifying EBV-
related Hodgkin lymphoma.
© American Society for Clinical Pathology
Downloaded from https://academic.oup.com/ajcp/article-abstract/117/2/259/1758556 by guest on 17 December 2019
Epidemiologic, molecular, and immunologic evidence
links Hodgkin lymphoma to Epstein-Barr virus (EBV) infec-
tion. A history of infectious mononucleosis is associated with
an increased risk of Hodgkin disease, and serologic studies
demonstrate high titers against viral antigens before tumor
diagnosis.1-3 EBV has been detected in the malignant Reed-
Sternberg/Hodgkin (RS/H) cells in approximately 40% of
patients with Hodgkin disease, but EBV-relatedness depends
on age, sex, geographic location, HIV status, and other char-
acteristics of the study population. 3-5 Of the 4 major
histopathologic subtypes of Hodgkin lymphoma, the mixed
cellularity subtype is most frequently EBV-associated (70%),
followed by lymphocyte depletion (50%), nodular sclerosis
(20%), and lymphocyte predominant subtypes (<5%).4,6
Accurate laboratory tests to detect EBV are needed for
purposes of basic and epidemiologic research and for clinical
management. From a clinical standpoint, tests for EBV can
be used to help establish a correct diagnosis in patients whose
histologic lesion has overlapping features of Hodgkin
lymphoma, anaplastic large cell lymphoma, or reactive
lymphoid hyperplasia.7 In addition, identification of tumor-
associated EBV implies that the patient is a candidate for
laboratory monitoring of tumor burden based on molecular
assays for EBV viral load8-10 and for treatment using EBV-
targeted oncologic therapies.11-13 Although somewhat unset-
tled, evidence from several studies suggests that EBV related-
ness is a prognostic factor and influences survival.8,14-19
Histochemical assays are widely used for localizing
Epstein-Barr viral nucleic acid or protein to the malignant
RS/H cells. These tumor cells express a limited spectrum of
viral proteins, including EBV nuclear antigen 1 and the latent
membrane proteins (LMP1, LMP2a, and LMP2b). In addition,
Am J Clin Pathol 2002;117:259-267 259
Gulley et al / EBER AND LMP1 INTERPRETATION GUIDELINES
EBV-encoded RNA (EBER1 and EBER2) is expressed
abundantly, although EBER transcripts are nonpolyadeny-
lated and remain untranslated. EBER in situ hybridization
has been recommended as the best test for detecting and
localizing latent EBV in tissue samples.20-25 EBER tran-
scripts are naturally abundant in latently infected cells, with
levels often exceeding 1 million copies per cell.26 Although
EBER in situ hybridization is widely used to define cases of
EBV-related Hodgkin disease, several pitfalls in technique
and interpretation have been described. For example, false-
positive EBER interpretations are attributable to confusion
related to latent infection of background lymphocytes,
nonspecific staining, or cross-reactivity with mucin, yeast, or
plant materials.20 False-negative results may be a conse-
quence of RNA degradation.27 To overcome this problem,
some laboratories use a control stain for a ubiquitous cellular
transcript to ensure that RNA is preserved and available for
hybridization in the cells of interest. The U6 cellular tran-
script is a particularly appropriate control because it is
similar to EBER in terms of size, abundance, and intranu-
clear distribution.28 Even when RNA is well preserved, some
investigators have described false-negative EBER results in
tumors shown to contain EBV by other methods or at
previous time points.29-31 One tantalizing explanation is the
hit-and-run hypothesis, which theorizes that EBV might
once have been present in a tumor, but that all or part of the
EBV genome subsequently was lost. 31,32 To facilitate
research into this hypothesis, it is important to define criteria
for interpreting analytic tests for EBV and to describe an
approach for handling focal EBER expression among RS/H
cells of the same tumor. To date, such guidelines are lacking.
An alternative strategy commonly used for identifying
EBV-related Hodgkin lymphoma is immunohistochemical
detection of viral LMP1. LMP1 is expressed in the cyto-
plasm and surface membrane of RS/H cells but is rarely
expressed in latently infected background lymphocytes of
Hodgkin lymphoma.33 LMP1 is, however, expressed in the
benign lymphocytes of infectious mononucleosis and in a
minute fraction of lymphocytes in healthy, remotely infected
viral carriers. 34,35 LMP1 immunostains are appealing
because they are more economic and rapid than EBER
hybridization and more easily incorporated into routine clin-
ical laboratories. LMP1 immunostains usually are performed
using commercial antibodies on fixed, paraffin-embedded
tissues.36 Results should be interpreted with caution since
false-positive staining is reported in poorly fixed tissues, in
cells of the nervous system, and in some uninfected
hematopoietic elements, including eosinophils and plasma
cells.20,36,37 False-negative immunostain results are more
frequent in decalcified tissues. In addition, failure to identify
EBV could occur if LMP1 is down-regulated, as is seen in
Burkitt lymphoma.33,38,39
260 Am J Clin Pathol 2002;117:259-267
Although histochemical assays for EBER and LMP1 are
widely used in research and clinical settings, assay reliability
has never been characterized, and literature providing tech-
nical and interpretive guidance is sparse. As part of a study
of interobserver agreement, we used evidence from interob-
server performance to develop guidelines for interpreting
EBER in situ hybridization and LMP1 immunohistochem-
ical stains in the setting of Hodgkin lymphoma. To accom-
plish this, Hodgkin lymphoma tissues from 40 patients were
stained in 2 separate testing laboratories and were indepen-
Downloaded from https://academic.oup.com/ajcp/article-abstract/117/2/259/1758556 by guest on 17 December 2019
dently interpreted by 4 hematopathologists (M.L.G., F.E.C.,
M.B., and R.B.M.), who then met jointly to review the cases,
discuss technical and interpretive pitfalls, and recommend
interpretive criteria.
Materials and Methods
We obtained 40 cases of Hodgkin lymphoma from the
archival paraffin-embedded tissue files of 2 of us (R.F.A. and
M.L.G.). These cases previously had been diagnosed as
Hodgkin lymphoma, subclassified using standard morpho-
logic criteria,7 and assessed for EBV status by EBER in situ
hybridization and/or LMP1 immunohistochemical analysis.
On this basis, the 40 cases were selected to include 20 cases
of nodular sclerosis (10 EBV-positive and 10 EBV-negative)
and 20 cases of mixed cellularity (10 EBV-positive and 10
EBV-negative) Hodgkin lymphoma. A single representative
block chosen from each case was used for all subsequent
histochemical studies. Paraffin sections were cut and distrib-
uted to 2 testing laboratories (Johns Hopkins University,
Baltimore, MD, and the University of Texas Health Science
Center at San Antonio) for EBER, U6, and LMP1 histo-
chemical stains using these laboratories’ routine analytic
procedures as described in the following sections. All studies
were carried out under the supervision of an institutional
review board.
EBER In Situ Hybridization Procedures
EBER in situ hybridization was performed on each case
at each of the 2 testing laboratories using digoxigenin-
labeled riboprobes produced from plasmid templates
obtained from one of us (R.F.A.).40 A step-by-step descrip-
tion of one of the testing protocols, which uses EBER1 ribo-
probes, has been published.41 The other laboratory’s proce-
dure differed in that a cocktail of EBER1 and EBER2
riboprobes was used, and the secondary structures of the
riboprobes were disrupted immediately before use by boiling
in hybridization buffer for 3 minutes followed by plunging in
ice. One laboratory hybridized overnight, whereas the other
hybridized for only 3 hours. To control for RNA preserva-
tion, both laboratories used the same type of riboprobe,
© American Society for Clinical Pathology

namely one targeting cellular U6 RNA. Both laboratories
used RNase to facilitate washing away unbound probe and a
detection system based on application of antibody to digoxi-
genin linked to alkaline phosphatase. One laboratory coun-
terstained with 0.5% eosin, whereas the other used methyl
green counterstain. Known cases of EBV-related Hodgkin
lymphoma were used as controls by both laboratories.
Immunohistochemical Detection of LMP1 Expression
LMP1 immunohistochemical stains were performed at
each of the 2 testing laboratories using similar methods.
Both laboratories used a commercial cocktail of 4 mono-
clonal antibodies against LMP1 (CS1-4, DAKO, Carpinteria
CA). Antigen availability was enhanced by pretreatment for
5 minutes with Proteinase K (DAKO) at room temperature.
Endogenous peroxidase was quenched with sodium azide
and hydrogen peroxide, and slides were incubated with anti-
body diluted at 1:200. Bound antibody was detected with
biotinylated secondary antibody followed by peroxidase-
labeled streptavidin and diaminobenzidine color reagent
(LSAB2/HRP System, DAKO). Automated instrumentation
(DAKO Autostainer, DAKO) was used to facilitate sequen-
tial application of reagents. One testing laboratory used
hematoxylin counterstain, whereas the other used methyl
green counterstain. Known cases of EBV-related Hodgkin
lymphoma were used as positive controls.
Data Collection
After staining, the 40 EBER/U6 slide sets and 40 LMP1
slide sets from each preparing laboratory (ie, 80 for each
assay), together with the corresponding H&E-stained slides,
were combined by the study epidemiologist (S.L.G.) into
assay-specific batches of 40 tumors. These batches were
distributed to the 4 hematopathologists so that each patholo-
gist independently interpreted each EBER stain and, at a
separate sitting, each LMP1 stain. Every slide was distrib-
uted for a second round of interpretations so that each of the
4 pathologists saw each of the 160 slides twice for a total of
1,280 interpretations. Each batch of 40 tumors was assem-
bled randomly to include both histologic subtypes, materials
from both laboratories, and both first and second reviews
(except for the first and last batches). Between reviews,
slides were relabeled to blind raters to review sequence and
preparing laboratory. This design enabled analysis of both
interobserver and intraobserver reliability of assay interpreta-
tion.
Each pathologist was asked to evaluate the EBER or
LMP1 slide in conjunction with an associated H&E-stained
slide and to interpret each result as positive, negative, or
uncertain with respect to the presence of EBV in RS/H cells.
The pathologist was asked to interpret a slide as positive if
chromogen was localized to any proportion of these cells.
© American Society for Clinical Pathology
Hematopathology / ORIGINAL ARTICLE
Negative staining was defined as the absence of chromogen
above background levels in the cells of interest. Each case
was rated as easy, intermediate, or difficult to interpret. The
pathologist also noted any factors that hampered interpreta-
tions, and these comments were recorded on data sheets
accompanying each shipment.
Development of Interpretive Guidelines
Following all reviews of each of the 40 Hodgkin cases, a
summary sheet was prepared listing the results of the 16
Downloaded from https://academic.oup.com/ajcp/article-abstract/117/2/259/1758556 by guest on 17 December 2019
interpretations and 16 difficulty ratings for the EBER and
LMP1 stains performed at each of the 2 testing laboratories.
These data, together with any pathologist comments, were
reviewed along with the slides at a consensus conference in
which interpretation guidelines were developed and tech-
nical and interpretive pitfalls were discussed. For this
purpose, all investigators assembled around a multiheaded
microscope to review the cases. For the first time, the pathol-
ogists were able to compare the EBER and LMP1 stains on a
given case, and they assigned a consensus interpretation on
the EBV relatedness of each case. Discrepancies in their
original interpretations were resolved, and the difficulty
ratings were discussed in conjunction with notes about
factors initially hindering interpretation. Based on these
discussions, guidelines for evaluating results were devel-
oped, and specific criteria for interpreting results were estab-
lished by consensus.
Data Analysis
Data were tabulated and analyzed by an epidemiologist
(S.L.G.) and a statistician (S.J.S.). For each of the 320 slides
reviewed, interpretation agreement among the 4 pathologists
was evaluated using the kappa statistic and associated 95%
confidence interval. A kappa greater than 0 indicates more
agreement than expected by chance; kappa is considered to
indicate fair to good agreement between 0.40 and 0.75, and
very good agreement above 0.75.42 For each of the 320
slides reviewed, a consensus difficulty rating among the 4
reviewers was assigned according to the following criteria:
The interpretation was considered easy if 3 or more patholo-
gists rated the interpretation as easy, difficult if 2 or more
pathologists rated it as difficult, and intermediate for all other
combinations. The kappa statistics stratified by difficulty
level were computed for interpretation agreement among the
4 pathologists.
Results
For all 160 EBER in situ hybridization reviews (ie, 40
from each laboratory, each reviewed twice), consensus inter-
pretation was considered easy in 70% of reviews, intermediate
Am J Clin Pathol 2002;117:259-267 261
Gulley et al / EBER AND LMP1 INTERPRETATION GUIDELINES

1.2 uncertain for both markers. When both markers were evalu-
ated in tandem, all 40 cases could be confidently assigned as
1.0
EBV-positive or EBV-negative. EBER and LMP1 were coex-
0.8 pressed in 17 cases, neither marker was expressed in 17
cases, 2 cases were uncertain for EBER and negative for
0.6
kappa

LMP1, 2 cases were uncertain for EBER and positive for

0.4 LMP1, and the remaining 2 cases were positive for EBER
and uncertain for LMP1. Discussion of the discrepancies led
0.2
to the conclusion that proper interpretation required the
0.0 ability to morphologically identify RS/H cells and to distin-

Downloaded from https://academic.oup.com/ajcp/article-abstract/117/2/259/1758556 by guest on 17 December 2019

guish them from background lymphocytes and other
– 0.2
nontumor cells. In addition, it was recommended that an
Easy Intermediate Difficult
H&E-stained section always be reviewed in parallel with
Difficulty of Interpretation each EBER or LMP1 stain, so that the prevalence, distribu-
❚Figure 1❚ Agreement among 4 pathologists interpreting 320 tion, and cytologic features of the RS/H cells on the H&E-
Epstein-Barr virus–encoded RNA and latent membrane protein stain section could guide assessment of RS/H cells on the
1 histochemical stains. Difficulty ratings are displayed on the x corresponding EBER- or LMP1-stained slide.
axis, and the corresponding kappa statistic and 95% confidence
interval (denoted by error bars) representing level of agreement EBER In Situ Hybridization
among the 4 pathologists are displayed on the y axis. The pathologists made the following observations about
EBER hybridizations and their interpretation. The EBER
❚Table 1❚ signal was localized to the nucleus ❚Image 1❚. Sometimes the
Agreement Among Four Pathologists Interpreting EBER and chromogen obliterated the entire nucleus, and other times the
LMP1 Histochemical Stains Stratified by Difficulty Ratings
staining was weaker and appeared to segregate with the chro-
Agreement: Multirater kappa matin, sparing the nucleolus. No cytoplasmic signal was seen
Analyte/Consensus Difficulty Rating* Statistic (95% CI) in tumor cells, except during mitosis where the nuclear
Epstein-Barr virus–encoded RNA membrane was dissolved.
Easy (n = 112) 0.982 (0.965 to 1.00) Before concluding that a slide is EBER-negative, it is
Intermediate (n = 39) 0.485 (0.331 to 0.645)
Difficult (n = 9) –0.091 (–0.223 to 0.073) essential to evaluate a control slide, run in parallel, demon-
Latent membrane protein 1 strating that RNA is present and available for hybridization
Easy (n = 117) 0.983 (0.964 to 1.00)
Intermediate (n = 30) 0.696 (0.579 to 0.825)
in the cells of interest. The U6 control RNA targeted in this
Difficult (n = 13) 0.157 (–0.006 to 0.356) study was localized to the nucleus of all types of cells
(Image 1B). In the experimental samples, RNA often was
CI, confidence interval.
* Interpretation was considered easy if 3 or more pathologists rated the interpretation only partially preserved, as evidenced by U6 signal in some
as easy, difficult if 2 or more pathologists rated it as difficult, and intermediate for
all other combinations.
but not all cells of a given field or in some but not all areas
of the section.
The EBER stains evaluated in the present study had a
in 24%, and difficult in 7%. For all 160 LMP1 immunostain purple chromogen and either a methyl green counterstain
reviews, consensus interpretation was considered easy in 73%, producing a green backdrop or an eosin counterstain
intermediate in 19%, and difficult in 8%. These ratings producing a pink backdrop (Images 1C and 1D). The methyl
suggest that difficulty of interpretation did not differ substan- green counterstain was easier to interpret for several reasons.
tially between EBER stains and LMP1 stains. There was First, purple and green contrasted better than did purple and
generally very good interobserver agreement among the 4 pink. Second, eosinophils were distracting on eosin stains
pathologists for EBER (kappa = 0.813; 95% confidence but not on methyl green stains. Third, eosin counterstain
interval, 0.767-0.859) and for LMP1 (kappa = 0.826; 95% sometimes resulted in “slivers” of crystalline stain precipitate
confidence interval, 0.782-0.871). Not surprisingly, the level that were not formed when methyl green was used. Finally,
of agreement tended to vary by consensus difficulty rating, as methyl green staining provided improved discrimination of
shown in ❚Figure 1❚ overall and in ❚Table 1❚ by assay. benign vs malignant cytologic features because methyl green
At the consensus conference, consensus interpretation of highlighted chromatin, whereas eosin primarily highlighted
the cases yielded 19 EBER-positive, 17 EBER-negative, and cytoplasmic proteins.
4 uncertain EBER results, and 19 LMP1-positive, 19-LMP1 The EBER signal varied from weak to strong in intensity.
negative, and 2 uncertain cases for LMP1. No cases were Intensity seemed to relate primarily to the degree of RNA

262 Am J Clin Pathol 2002;117:259-267 © American Society for Clinical Pathology

 Hematopathology / ORIGINAL ARTICLE

A B

Downloaded from https://academic.oup.com/ajcp/article-abstract/117/2/259/1758556 by guest on 17 December 2019

C D

❚Image 1❚ Epstein-Barr virus–encoded RNA (EBER) in situ hybridization of Hodgkin tumors. A, Purple EBER signal is localized to
the nucleus of a Reed-Sternberg/Hodgkin cell. Multiple black pigment deposits represent fixation artifact. B, U6 control
hybridization confirms RNA preservation in the tissue as evidenced by U6 expression in the nuclei of reactive and malignant
cells alike. C, A single small lymphoid cell in the center expresses the purple EBER signal, while the remaining nucleated cells
stain only with methyl green counterstain. D, A single small lymphoid cell in the center expresses the purple EBER signal,
while the remaining cells stain pink owing to the eosin counterstain, including bright pink eosinophils. (×150)

preservation. It was impossible to assess the extent to
which natural levels of EBER varied from cell to cell or
case to case. One of the most difficult aspects of inter-
preting EBER stains was distinguishing background
lymphocytes from RS/H cells. Rare (<0.1%), small, EBER-
positive cells with scant cytoplasm and clumped chromatin
were seen in many of the cases, and these were interpreted
as latently infected lymphocytes that might be present in
any viral carrier. More difficult to interpret were occasional
medium-sized lymphocytes with visible nucleoli expressing
EBER in a tumor in which the more obvious RS cells were
EBER-negative. To distinguish mononuclear RS/H cells
from latently infected benign lymphocytes, the pathologists
relied on the small size of the nucleolus, scant cytoplasm,
and lack of “lacunae” surrounding benign lymphocytes. In
questionable cases, it is important to carefully scan the slide
for EBER positivity in the more obvious RS cells having
large, lobated nuclei, prominent nucleoli, and more abun-
dant cytoplasm. The morphologist should be cautious in
interpreting a case as EBER-positive unless EBER is
expressed in at least 1 unequivocal tumor cell. As discussed
in the following section, immunohistochemical detection of

© American Society for Clinical Pathology Am J Clin Pathol 2002;117:259-267 263

Gulley et al / EBER AND LMP1 INTERPRETATION GUIDELINES

A B

Downloaded from https://academic.oup.com/ajcp/article-abstract/117/2/259/1758556 by guest on 17 December 2019

C D

❚Image 2❚ Latent membrane protein 1 (LMP1) immunohistochemical stains of Hodgkin tumors. A, Brown, granular LMP1 signal
is localized to the cytoplasm and surface membrane of a Reed-Sternberg cell. B, A mononuclear Reed-Sternberg/Hodgkin cell
expresses LMP1 in the upper right, several smaller eosinophils exhibit false-positive cytoplasmic staining, and black carbon
pigment deposits are seen in the lower left. C, Occasional plasma cells exhibit false-positive cytoplasmic staining with the CS1-
4 cocktail of antibodies used to identify LMP1. D, Nonspecific cytoplasmic staining in tumor cells and nontumor cells alike along
the edge of the slide is the so-called edge artifact. (Methyl green counterstain, ×150)

LMP1 expression is quite helpful for resolving cases with
poor RNA preservation or equivocal EBER results.
LMP1 Immunohistochemical Findings
Microscopic visualization of LMP1 immunostains
revealed that LMP1 was localized to the cytoplasm and
surface membrane of RS/H cells ❚Image 2❚. In some cells, the
LMP1 signal was concentrated in one segment of the surface
membrane to form a “cap.” LMP1 was not identified in
benign-appearing lymphocytes. Intensity of LMP1 staining in
RS/H cells varied from strong to weak. It is important to note
that the chromogen had a distinctly granular character. In
contrast, a faint homogeneous blush of cytoplasmic color was
nonspecific and could be seen in fibroblasts as well as apop-
totic or “mummified” tumor cells.
False-positive cytoplasmic signal was seen in
eosinophils, which often are quite numerous in classic
Hodgkin tissues. Eosinophils could be readily distinguished
from RS/H cells based on cell size and nuclear size and
shape and by the different character of cytoplasmic granu-
larity (Image 2B).
False-positive cytoplasmic LMP1 signal was evident in
some but not all plasma cells, suggesting cross-reactivity of
the antibody cocktail with certain plasma cell antigens.

264 Am J Clin Pathol 2002;117:259-267 © American Society for Clinical Pathology


Plasma cells were readily distinguished from RS/H cells based
on cell size, nuclear size and contours, and chromatin pattern
(Image 2C). In some tissues, particularly spleen, hemosiderin-
laden macrophages contained brown pigment that must not be
confused with LMP1 signal. Presumably carbon, melanin, or
other pigments could likewise mimic the chromogen.
Matching the cytomorphologic features and distribution
of the RS/H cells on the H&E stain to those on the LMP1
stain should help ensure that the correct cell population is
evaluated. A positive control slide is essential for demon-
strating that the system is working and to serve as a gauge of
the expected color and granular character of true-positive
staining. Likewise, a negative control stain in which the anti-
body is omitted should help assess nonspecific staining.
Focal Variation in EBER or LMP1 Expression Within a
Given Tissue Sample
When EBER was expressed in any RS/H cells, it usually
was expressed in all RS/H cells. However, in some cases, only
a proportion of RS/H cells were positive. The most common
reason for focal EBER staining seemed to be focal RNA
preservation. For this reason, it is important to evaluate the
corresponding U6 stain to identify areas where RNA was well
preserved and available for hybridization in the RS/H cells. If
EBER analysis is restricted to areas with adequate RNA preser-
vation, then EBER is nearly always expressed in an “all-or-
nothing” manner in the RS/H cell population of a given tumor.
When LMP1 was expressed in any RS/H cells, it usually
was expressed in all RS/H cells. Focal LMP1 was an alter-
nate outcome, occurring in 6 (15%) of 40 Hodgkin cases, or
6 (30%) of 20 LMP1-positive cases. In some of these cases,
fewer than 5% of tumor cells expressed LMP1.
“Edge artifact,” whereby the edges of the tissue had a
variably wide rim of nonspecific staining, was seen in some
of the LMP1 immunostains but in none of the EBER
hybridizations (Image 2D). LMP1 and EBER stains occa-
sionally had large (30-100 µm) spots of nonspecific staining
that we speculated were a consequence of either fixation arti-
fact (especially with B-5 fixative) or bubbles in the staining
reagents. Such zones of nonspecific staining were localized
to clusters of cells regardless of their histogenesis.
Discussion
Previous molecular studies revealed monoclonal EBV
DNA in Hodgkin tumor tissue, suggesting that the virus was
present before expansion of the malignant clone.43 The role
of EBV in disease pathogenesis and the prognostic and ther-
apeutic value of EBV infection remain subjects of ongoing
investigation. To facilitate basic and clinical research, it is
important to define criteria for interpreting the histochemical
© American Society for Clinical Pathology
Hematopathology / ORIGINAL ARTICLE
❚Table 2❚
Guidelines for Interpreting Epstein-Barr Virus–Encoded RNA
(EBER) In Situ Hybridization and Latent Membrane Protein 1
(LMP1) Immunohistochemical Stains in Hodgkin Lymphoma
• The morphologist must be competent in distinguishing Reed-
Sternberg/Hodgkin (RS/H) cells from nontumor cells. The
cytologic features and distribution of RS/H cells should be
assessed on matched H&E-stained sections before interpreting
EBER or LMP1 results.
• To interpret a case as positive, the EBER or LMP1 signal must
be unequivocally localized to RS/H cells. The fraction of RS/H
cells expressing EBER and LMP1 varies among cases, with
most cases having a high fraction of positive tumor cells. For
Downloaded from https://academic.oup.com/ajcp/article-abstract/117/2/259/1758556 by guest on 17 December 2019
purposes of identifying all Epstein-Barr virus–related Hodgkin
cases, just 1 unequivocal RS/H cell expressing EBER or LMP1
should be considered sufficient to call a case positive. Equivocal
cases frequently are resolved when EBER, LMP1, and H&E
stains are evaluated in parallel.
• The EBER signal is localized to the nucleus, sometimes sparing
or rimming the nucleus. A negative EBER result can be
interpreted as negative only if RNA is shown to be preserved
and available for hybridization in tumor cells. A variable
proportion of background lymphocytes express EBER (usually
0%-1%), and these must be distinguished from RS/H cells.
• The LMP1 signal is localized to the cytoplasm and surface
membrane and has a granular character. False-positive
cytoplasmic staining is seen in eosinophils and a fraction of
plasma cells. “Edge artifact” and other “background staining”
may contribute to false-positive signal in benign and malignant
cells alike.
• As with any histochemical stain, quality control measures should
be used to ensure adequacy of staining, preservation of
cytomorphologic features, and interpretive competency.
stains that are most commonly used to identify EBV in
Hodgkin lymphoma.
In the present study of 40 Hodgkin tumors evaluated by
4 expert hematopathologists, EBER stains were considered
interpretable in 36 cases and LMP1 stains in 38 cases. By
using the 2 stains in tandem, EBV status was assigned in all
40 cases. Testing and interpretation of these cases provided a
framework for developing guidelines for interpreting EBER
and LMP1 histochemical stains in Hodgkin tissue samples,
as summarized in ❚Table 2❚. A critical factor in interpreting
stain results is the ability to morphologically identify RS/H
cells and to distinguish them from background lymphocytes,
histiocytes, and other nontumor cells. We recommend that an
H&E-stained section be reviewed along with the histochem-
ical stain so as to match the prevalence, distribution, and
cytologic features of the RS/H cells on the H&E-stained
section to those on the corresponding EBER- or LMP1-
stained slide.
In our experience, EBER and LMP1 usually were
expressed in an all-or-nothing manner among the RS/H cells
of a given case. However, in some cases, only a proportion of
RS/H cells were positive for these markers. The most
common reason for focal EBER staining seemed to be focal
RNA preservation. For this reason, it is important to evaluate
the corresponding control hybridization to identify areas in
Am J Clin Pathol 2002;117:259-267 265
Gulley et al / EBER AND LMP1 INTERPRETATION GUIDELINES
which RNA was well preserved and available for hybridiza-
tion in the cells of interest. To avoid a technical pitfall that
might lead to false-negative EBER results, it is important that
the EBER signal be at least as strong as the control signal so
that control gene expression signifies that enough RNA is
preserved to allow detection of any EBER-positive cells.
Cases in which RNA is lacking in the RS/H cells cannot
be interpreted as negative for EBER, since it is likely that the
RNA has been destroyed at some point during tissue prepara-
tion. To minimize RNA degradation during sectioning, we
recommend the use of a sterile water bath to float sections and
the wearing of gloves so that the slides are never touched with
bare hands. Since treatment with RNase is an integral part of
the posthybridization wash procedures, the RNase solutions
must not be allowed to contaminate any of the reagents or
surfaces to which slides are exposed before hybridization. At
the time of tissue procurement, tissues should be sliced and
fixed as soon as possible, but not overfixed before processing.
Formalin and B-5 are both suitable fixatives. Fixation artifact
often is seen with B-5 fixative and sometimes with formalin
fixation, resulting in black spider-like precipitants that inter-
fere with visualization of the intended chromogen. Decalcifi-
cation does not necessarily interfere with EBER or LMP1
staining, and appropriate control stains can be done to assess
the ability to detect similar targets known to be expressed in
these types of tissues. As always, small biopsy specimens and
those with few RS/H cells are fraught with concerns about
sampling error.
When only a fraction of atypical cells express EBER or
LMP1, it often is difficult to be certain that these are RS/H
cells. For this reason, it is important that the microscopist be
well versed in the morphologic diagnosis of Hodgkin
lymphoma and that the tissue sample be of adequate size and
quality to permit accurate interpretation. In the present study,
LMP1 expression was focal in about 15% of cases. Because
staining could be focal and weak, we consider it important to
scan the entire slide using a medium to high-power objective.
Otherwise, weak, focal LMP1 staining could be overlooked.
The biologic explanation for variation in LMP1 expression
levels is not well understood. Down-regulation of LMP1
could be a function of the degree of tumor cell differentiation
or of the immune system’s ability to recognize and destroy
cells expressing foreign antigens. Further research is
required to define the factors influencing LMP1 expression
levels. For purposes of identifying all EBV-related Hodgkin
cases, we propose that just 1 unequivocal RS/H cell
expressing EBER or LMP1 should be considered sufficient
to call a tumor positive for EBV. Equivocal cases are
frequently resolved when EBER, LMP1, and H&E stains are
evaluated in parallel. When many EBER-positive small
lymphocytes are present, it is wise to consider infectious
mononucleosis in the differential diagnosis.
266 Am J Clin Pathol 2002;117:259-267
In the present study, Proteinase K was used to facilitate
antigen or nucleic acid retrieval before application of anti-
body or probe. It is important that Proteinase K digestion be
sufficient to achieve this objective without destroying the
cytomorphologic features of the tissue. Individual tumors
may benefit from different digestion conditions, and it may
be necessary to adjust the Proteinase K concentration or
incubation time to optimize stain quality.
A number of technical and interpretive issues influence
the outcome of EBER and LMP1 histochemical tests applied
Downloaded from https://academic.oup.com/ajcp/article-abstract/117/2/259/1758556 by guest on 17 December 2019
to Hodgkin tumors. The guidelines provided herein,
combined with proper attention to the quality control
measures that guide all histochemical procedures, should
facilitate uniform interpretation of these tests. Because
EBER and LMP1 stains are each subject to false-negative
interpretation, we recommend that future studies of EBV-
related Hodgkin lymphoma rely on EBER and LMP1 histo-
chemical stains in concert when determining the EBV relat-
edness of a particular Hodgkin tumor.
From the 1Department of Pathology, University of Texas Health
Science Center at San Antonio; 2Northern California Cancer
Center, Union City, CA; 3Department of Pathology, University of
Pittsburgh, Pittsburgh, PA; and 4Johns Hopkins University,
Baltimore, MD.
Funded by a grant R03 CA77125 from the National Cancer
Institute, Bethesda, MD.
Address reprint requests to Dr Gulley: Dept of Pathology,
University of North Carolina, Chapel Hill, NC 27599-7525.
Acknowledgments: We thank Phyllis Eagan, Alexandra
Morgan, Patricia Harasty, and Susan Stewart, PhD, for their
contributions to this study.
References
1. Hjalgrim H, Askling J, Sorensen P, et al. Risk of Hodgkin’s
disease and other cancers after infectious mononucleosis.
J Natl Cancer Inst. 2000;92:1522-1528.
2. Mueller N, Evans A, Harris NL, et al. Hodgkin’s disease and
Epstein-Barr virus: altered antibody pattern before diagnosis.
N Engl J Med. 1989;320:689-695.
3. Mueller N. Hodgkin’s disease. In: Scottenfeld D, Fraumeni JF
Jr, eds. Cancer Epidemiology and Prevention. New York, NY:
Oxford University Press; 1996:893-919.
4. Glaser SL, Lin RJ, Stewart SL, et al. Epstein-Barr
virus–associated Hodgkin’s disease: epidemiologic
characteristics in international data. Int J Cancer.
1997;70:375-382.
5. Siebert JD, Ambinder RF, Napoli VM, et al. Human
immunodeficiency virus–associated Hodgkin’s disease
contains latent, not replicative, Epstein-Barr virus. Hum
Pathol. 1995;26:1191-1195.
6. von Wasielewski R, Werner M, Fischer R, et al. Lymphocyte-
predominant Hodgkin’s disease: an immunohistochemical
analysis of 208 reviewed Hodgkin’s disease cases from the
German Hodgkin Study Group. Am J Pathol. 1997;150:793-803.
© American Society for Clinical Pathology

7. Harris NL. Hodgkin’s disease: classification and differential
diagnosis. Mod Pathol. 1999;12:159-175.
8. Gallagher A, Armstrong AA, MacKenzie J, et al. Detection of
Epstein-Barr virus (EBV) genomes in the serum of patients
with EBV-associated Hodgkin’s disease. Int J Cancer.
1999;84:442-448.
9. Lei KI, Chan LY, Chan WY, et al. Quantitative analysis of
circulating cell-free Epstein-Barr virus (EBV) DNA levels in
patients with EBV-associated lymphoid malignancies. Br J
Haematol. 2000;111:239-246.
10. Drouet E, Brousset P, Fares F, et al. High Epstein-Barr virus
serum load and elevated titers of anti-ZEBRA antibodies in
patients with EBV-harboring tumor cells of Hodgkin’s disease.
J Med Virol. 1999;57:383-389.
11. Rooney CM, Roskrow MA, Suzuki N, et al. Treatment of
relapsed Hodgkin’s disease using EBV-specific cytotoxic T
cells. Ann Oncol. 1998;9(suppl 5):S129-S132.
12. Su Z, Peluso MV, Raffegerst SH, et al. The generation of
LMP2a-specific cytotoxic T lymphocytes for the treatment of
patients with Epstein-Barr virus–positive Hodgkin disease.
Eur J Immunol. 2001;31:947-958.
13. Roskrow MA, Suzuki N, Gan Y, et al. Epstein-Barr virus
(EBV)-specific cytotoxic T lymphocytes for the treatment of
patients with EBV-positive relapsed Hodgkin’s disease. Blood.
1998;91:2925-2934.
14. Armstrong AA, Lennard A, Alexander FE, et al. Prognostic
significance of Epstein-Barr virus association in Hodgkin’s
disease. Eur J Cancer. 1994;7:1045-1046.
15. Vestlev PM, Pallesen G, Sandvej K, et al. Prognosis of
Hodgkin’s disease is not influenced by Epstein-Barr virus
latent membrane protein. Int J Cancer. 1992;50:670-671.
16. Morente MM, Piris MA, Abraira V, et al. Adverse clinical
outcome in Hodgkin’s disease is associated with loss of
retinoblastoma protein expression, high Ki67 prolifer-
ation index, and absence of Epstein-Barr virus–latent
membrane protein 1 expression. Blood. 1997;90:2429-
2436.
17. Murray PG, Billingham LJ, Hassan HT, et al. Effect of
Epstein-Barr virus infection on response to chemotherapy and
survival in Hodgkin’s disease. Blood. 1999;94:442-447.
18. Clarke CA, Glaser SL, Dorfman RF, et al. Epstein-Barr virus
and survival after Hodgkin disease in a population-based series
of women. Cancer. 2001;91:1579-1587.
19. Glavina-Durdov M, Jakic-Razumovic J, Capkun V, et al.
Assessment of the prognostic impact of the Epstein-Barr
virus–encoded latent membrane protein-1 expression in
Hodgkin’s disease. Br J Cancer. 2001;84:1227-1234.
20. Gulley ML. Molecular diagnosis of Epstein-Barr virus–related
diseases. J Mol Diagn. 2001;3:1-10.
21. Ambinder RF, Mann RB. Detection and characterization of
Epstein-Barr virus in clinical specimens. Am J Pathol.
1994;145:239-252.
22. Brousset P, Meggetto F, Chittal S, et al. Assessment of the
methods for the detection of Epstein-Barr virus nucleic acids
and related gene products in Hodgkin’s disease. Lab Invest.
1993;69:483-490.
23. Hamilton-Dutoit SJ, Pallesen G. Detection of Epstein-Barr
virus small RNAs in routine paraffin sections using non-
isotopic RNA/RNA in situ hybridization. Histopathology.
1994;25:101-111.
24. Armstrong AA, Weiss LM, Gallagher A, et al. Criteria for the
definition of Epstein-Barr virus association in Hodgkin’s
disease. Leukemia. 1992;6:869-874.
© American Society for Clinical Pathology
Hematopathology / ORIGINAL ARTICLE
25. Wu TC, Mann RB, Charache P, et al. Detection of EBV gene
expression in Reed-Sternberg cells of Hodgkin’s disease. Int J
Cancer. 1990;46:801-804.
26. Clemens MJ. The small RNAs of Epstein-Barr virus. Mol Biol
Rep. 1993;17:81-92.
27. Arber JM, Weiss LM, Chang KL, et al. The effect of
decalcification on in situ hybridization. Mod Pathol.
1997;10:1009-1014.
28. Andersen J, Zieve GW. Assembly and intracellular transport
of snRNP particles. Bioessays. 1991;13:57-64.
29. Srinivas SK, Sample JT, Sixbey JW. Spontaneous loss of viral
episomes accompanying Epstein-Barr virus reactivation in a
Downloaded from https://academic.oup.com/ajcp/article-abstract/117/2/259/1758556 by guest on 17 December 2019
Burkitt’s lymphoma cell line. J Infect Dis. 1998;177:1705-
1709.
30. Delecluse HJ, Marafioti T, Hummel M, et al. Disappearance of
the Epstein-Barr virus in a relapse of Hodgkin’s disease.
J Pathol. 1997;182:475-479.
31. Razzouk BI, Srinivas S, Sample CE, et al. Epstein-Barr virus
DNA recombination and loss in sporadic Burkitt’s lymphoma.
J Infect Dis. 1996;173:529-535.
32. Sixbey JW. Epstein-Barr virus DNA loss from tumor cells and
the geography of Burkitt’s lymphoma. EBV Rep. 2000;7:37-40.
33. Hummel M, Anagnostopoulos I, Dallenbach F, et al. EBV
infection patterns in Hodgkin’s disease and normal lymphoid
tissue: expression and cellular localization of EBV gene
products. Br J Haematol. 1992;82:689-694.
34. Reynolds DJ, Banks PM, Gulley ML. New characterization of
infectious mononucleosis and a phenotypic comparison with
Hodgkin’s disease. Am J Pathol. 1995;146:379-388.
35. Babcock GJ, Thorley-Lawson DA. Tonsillar memory B cells,
latently infected with Epstein-Barr virus, express the restricted
pattern of latent genes previously found only in Epstein-Barr
virus–associated tumors. Proc Natl Acad Sci U S A.
2000;97:12250-12255.
36. Jiwa NM, Oudejans JJ, Dukers DF, et al. Immunohisto-
chemical demonstration of different latent membrane
protein-1 epitopes of Epstein-Barr virus in lymphoproliferative
diseases. J Clin Pathol. 1995;48:438-442.
37. Hammer RD, Scott M, Shahab I, et al. Latent membrane
protein antibody reacts with normal hematopoietic precursor
cells and leukemic blasts in tissues lacking Epstein-Barr virus
genome by polymerase chain reaction. Am J Clin Pathol.
1996;106:469-474.
38. Tao Q, Robertson KD, Manns A, et al. Epstein-Barr virus
(EBV) in endemic Burkitt’s lymphoma: molecular analysis of
primary tumor tissue. Blood. 1998;91:1373-1381.
39. Niedobitek G, Agathanggelou A, Rowe M, et al. Hetero-
geneous expression of Epstein-Barr virus latent proteins in
endemic Burkitt’s lymphoma. Blood. 1995;86:659-665.
40. Barletta JM, Kingma DW, Ling Y, et al. Rapid in situ
hybridization for the diagnosis of latent Epstein-Barr virus
infection. Mol Cell Probes. 1993;7:105-109.
41. Fan H, Gulley ML. Molecular methods for detecting Epstein-
Barr virus. In: Killeen AA, ed. Molecular Pathology Protocols.
Totowa, NJ: Humana Press; 2001.
42. Posner KL, Sampson PD, Caplan RA, et al. Measuring
interrater reliability among multiple raters: an example of
methods for nominal data. Stat Med. 1990;9:1103-1115.
43. Gulley ML, Eagan PA, Quintanilla-Martinez L, et al. Epstein-
Barr virus DNA is abundant and monoclonal in the Reed-
Sternberg cells of Hodgkin’s disease: association with mixed
cellularity subtype and Hispanic American ethnicity. Blood.
1994;83:1595-1602.
Am J Clin Pathol 2002;117:259-267 267