Expression of multiple human endogenous retrovirus surface envelope proteins
in ovarian cancer
Feng Wang-Johanning1,2*, Jinsong Liu3, Kiera Rycaj1, Miao Huang1, Kate Tsai1, Daniel G. Rosen3, Dung-Tsa Chen4,
Danielle W. Lu5, Kirstin F. Barnhart1and Gary L. Johanning1
1Department of Veterinary Sciences and Michale E. Keeling Center for Comparative Medicine and Research,
University of Texas MD Anderson Cancer Center, Houston, TX
2Department of Immunology, University of Texas MD Anderson Cancer Center, Houston, TX
3Department of Pathology, University of Texas M.D. Anderson Cancer Center, Houston, TX
4Biostatistics and Bioinformatics Unit, University of Alabama at Birmingham, Birmingham, AL
5Department of Pathology, Huntington Memorial Hospital, Pasadena, CA
Individual classes of human endogenous retrovirus (HERV) genes
and proteins are expressed in cancer, but expression of more than
one type of HERV is rare. We report here the expression of multiple
HERV genes and proteins in ovarian cell lines and tissues. Expres-
sion of HERV-K env mRNA was greater in ovarian epithelial
tumors than in normal ovarian tissues (N 5 254). The expression of
this protein on the surface and in the cytoplasm of ovarian cancer
cells was confirmed using anti-HERV-K specific antibody by flow
cytometric analysis. The frequency of expression of HERV-K env
protein in multitissue microarrays (N 5 641) was determined by
immunohistochemistry and a significant correlation with tumor his-
totype was found. A significantly increased expression of HERV-K
was observed in tumors with low malignant potential and low grade,
relative to expression in normal ovarian tissues. The increase in
expression of HERV-K env protein took place in a stepwise fashion
in serous papillary adenocarcinoma. Interestingly, we found that
other classes of HERV env mRNAs, including ERV3 and HERV-E,
are expressed in the same ovarian cancer tissues that expressed
HERV-K. Furthermore, anti-HERV antibodies including anti-
ERV3 (30%), anti-HERV-E (40%) and anti-HERV-K (55%) were
detected in patients with ovarian cancer, but not in normal female
controls. HERV env proteins are frequently transcribed and trans-
lated in ovarian epithelial tumors, and multiple HERV families are
detectable in ovarian cancer. HERV env proteins, and especially
those expressed on the cell surface, may serve as novel tumor targets
for detection, diagnosis and immunotherapy of ovarian cancer.
' 2006 Wiley-Liss, Inc.
Key words: ovarian cancer; human endogenous retroviruses; surface
envelope proteins; anti-HERV antibodies; tumor targets
Ovarian cancer represents a great clinical challenge in gyneco-
logic oncology. The incidence of ovarian cancer is highest in the
United States and is the 5th most common malignancy among U.S.
women. There will be 22,220 estimated new cases of ovarian carci-
noma diagnosed in 2005, and 16,210 women die of the disease
annually, according to the American Cancer Society. Approxi-
mately 90% of all ovarian cancers are epithelial types. Nonepithelial
types include the sex cord-stromal tumors (6%), germ-cell tumors
(3%) and indeterminate tumors (1%).
The etiology of ovarian cancer is poorly understood. Tumor
stage at diagnosis is an important prognostic factor, and since most
ovarian tumors are not symptomatic until an advanced stage,
efforts to develop screening programs for ovarian cancer have
been undertaken. Unfortunately, the value of screening for ovarian
cancer has not been clearly established.
Human endogenous retroviruses (HERVs) and elements contain-
ing long terminal repeat (LTR)-like sequences may comprise up to
8% of the human genome.1HERVs are believed to have germline-
integrated their DNA into the human genome over 30 million years
ago. Former exogenous retroviruses may have infected the human
germ line; alternatively, infective exogenous retroviruses may have
been derived from integrated HERV sequences. HERVs contain
over 200 distinct groups and subgroups.2HERVs have been divided
into 3 broad classes3: Class 1 or g-retroviruses include HERV-E
and ERV3; class II or b-retroviruses include HERV-K and MMTV
and class III HERVs are related to spumaretroviruses. HERVs have
been grouped into classes based on sequence similarities, but the
various systems of HERV classification are somewhat confusing
and no classification scheme is universally accepted. The accumula-
tion of mutations has led to a loss of infectivity of HERVs, and in
general they are largely noninfectious retroviral remnants. However,
open reading frames (ORFs) have been observed for ERV3,4
and others10; 3 of them are functional, namely HERV-K108,6
The ERV3 retrovirus consists of a single-copy, full-length provi-
rus that has a long ORF in the env region,11,12functional LTRs and
a nondefective env glycoprotein gene.13Eight polymorphic sites
have been found in ERV3, and 1 of them is a defective allele with a
premature stop codon.14ERV3 env transcripts are present in pla-
centa and tumor cell lines,15–18and its expression may be involved
in human vascular disease.19The gag-pol sequences of the 8.8 kb
HERV-E proviral gene, which is a human C-type retrovirus, are
?40% homologous to Moloney murine leukemia virus (MuLV)
DNA. HERV-E 4-1 mRNA expression has been observed in human
placenta,20breast carcinoma,21colon carcinoma, germ cell tumors
and prostate adenocarcinoma.22HERV-K was originally identified
by its homology to the mouse mammary tumor virus, and is tran-
scriptionally active in several human cancer tissues23–25as well as
tumor cell lines, most notably the human breast cancer cell line
T47D,26,27and breast cancer tissues, as reported by Willer et al.28
and our group.29HERV-K retrovirus-like particles have been de-
scribed that encode for viral particles in the teratocarcinoma cell
line GH,23–25and the breast cancer cell line T47D.30The K family
encodes the regulatory protein Rec (formerly designated cORF)
and is comparable to HIV-1 Rev protein.25,31–35
We previously reported that transcripts of HERV-K with coding
potential for the env region of the gene are expressed frequently in
human breast cancer,29,36and that HERV-E and ERV3 are ex-
pressed in prostate cancer.37We found that the expression of HERV
env transcripts in breast cancers was enhanced by hormone stimula-
tion, and was specifically associated with the progression of neopla-
sia.36In the current study, we evaluated HERV expression in ovar-
ian cancer cells and tissues, and report that, in contrast to most other
cancers, which express 1 or at most 2 HERVs, at least 3 HERV env
genes are expressed in ovarian cancer.
Grant sponsor: Susan G. Komen Breast Cancer Foundation; Grant
number: BCTR0402892; Grant sponsor: M.D. Anderson Cancer Center.
*Correspondence to: Departments of Veterinary Sciences and Immu-
nology, The University of Texas M.D. Anderson Cancer Center, 650 Cool
Water Drive, Bastrop, TX 78602, USA. Fax: 1512-332-5218.
Received 24 May 2006; Accepted 18 July 2006
Int. J. Cancer: 120, 81–90 (2006)
' 2006 Wiley-Liss, Inc.
Publication of the International Union Against Cancer
Material and methods
Cells and tissues
The human ovarian surface epithelial cancer cell lines SKOV3,
OVCA 430, OVCA 433, OVCA 420, OVCAR3, DOV 13 and
OVCA 429, and the normal human ovarian epithelial cell lines
NOE 114, NOE 116, NOE 113 and NOE 119 were gifts from Dr.
Robert C. Bast Jr., University of Texas M. D. Anderson Cancer
Center. Cells were cultured in the media recommended by this
laboratory. The immortalized normal human ovarian epithelial cell
lines T29, T72 and T80 were generated from human ovarian sur-
face epithelial cells that had been described previously.38Tissue
samples were snap-frozen and stored at 280?C until RNA isola-
tion. For in situ hybridization and immunohistochemistry, forma-
lin-fixed, paraffin-embedded tissues were used.
Oligonucleotide primers derived from the sequences encoding
the env surface proteins of HERV-K, ERV3 and HERV-E were
used to amplify cDNA prepared from human ovarian tissues and
cell lines as described previously.29In our study, a sense primer
(nucleotide [nt] 6674-6698; Accession number: AF07408639) spe-
cific for type 2 HERV-K env genes was also used to detect Type 2
HERV-K env mRNA transcripts, as described previously.36Previ-
ously described primer pairs were used to amplify env reading
frame transcripts that include np9.40
RT-PCR and real-time RT-PCR
RNA was prepared and total RNA was reverse-transcribed and
amplified as described previously.29The cDNA was analyzed
using primers that recognize human b-actin to confirm equivalent
loading. One-step RT-PCR was performed to quantitate the
expression of the HERV-K env gene in various ovarian specimens
as described previously.36Homo sapiens ribosomal protein S9
(GenBank accession number XM 008957.2) was used as an en-
dogenous control.41Linear extrapolation of the cycle threshold
(CT) values of HERV-K were obtained from ovarian specimens,
and then divided by the relative amounts of S9, which were quan-
titated by linear extrapolation from the CTvalues of the same
Synthesis of HERV env fusion proteins and production
of anti-HERV env protein antibodies
HERV cDNAs obtained from ovarian cancer tissues were
cloned into the corresponding enzyme-digested QIA expression
vector (pQE30; Qiagen, Valencia, CA), which contains a 6-His
tag at the N-terminus. Several recombinant fusion proteins were
produced in our laboratory including HERV-K Type 1 surface env
protein, HERV-E surface env protein, HERV-K gag protein,
ERV3 env protein and HERV-K plus surface protein, which is
Type 2 env surface proteins with a 292 bp insert (also designated
HERV-K env Type 2). These fusion proteins were further charac-
terized by sequencing, followed by induction of protein expression
and purification by affinity chromatography. These purified HERV
env fusion proteins were used to immunize rabbits for production
of polyclonal, or mice for production of monoclonal, anti-HERV
env protein antibodies using standard techniques.
Flow cytometry and immunoflorescence staining
Cultured cells were incubated with anti-HERV-K antibody
(5,691; 1:200 dilution) at 4?C for 30 min, followed by anti-rabbit
IgG-FITC secondary mAb (1:1,000 dilution) at 4?C for 15 min.
For staining in permeabilized conditions, cells were treated with
0.1% Triton X-100 in PBS, then incubated with primary antibody,
followed by secondary antibody. Samples were analyzed on a BD
FACSCaliburTMsystem (BD Biosciences, San Diego, CA). For
controls, samples were incubated with anti-rabbit IgG-FITC sec-
ondary mAb. These cells were subjected to fluorescence micros-
copy to confirm the expression of HERV-K surface protein.
Immunohistochemistry of ovarian tissue slides
Paraffin-embedded ovarian tissue specimens were cut into serial
5 lm sections, melted, deparaffinized in xylene, rehydrated in
ethanol and then fixed in 4% paraformaldehyde. The slices were
incubated with anti-HERV-K env polyclonal antibody (1:200 dilu-
tion), followed by incubation with anti-rabbit IgG biotin conjugate
antibody (1:1,000 dilution) and finally with ABC (ABC kit, Vector
Laboratories, Burlingame, CA) as described by the manufacturer.
development. Slices were then counterstained with hematoxylin.
Tissue microarray slides
Multiple tissue microarray (TMA) slide TMA1, containing 72
ovarian tissues from patients with various ovarian diseases, was
obtained from US Biomax (Rockville, MD). Slide TMA2, which
contains 85 ovarian tissues that included normal, mucinous cyst,
low malignant potential (LMP) and low-grade (LG) and high-grade
(HG) carcinomas, was obtained from The University of Texas M.
D. Anderson Cancer Center Department of Pathology. TMA3 that
contains 484 cases of various ovarian cancer tissues with clinical
follow-up information was also obtained from The University of
Texas M. D. Anderson Cancer Center Department of Pathology.
Immunohistochemistry of multiple TMA slides
Immunohistochemistry was performed on multiple TMA slides
using a DAKO autostainer universal staining system. These multi-
ple TMA slides provided us with a means by which to compare
the expression of HERV-K env SU protein in multiple tissues
under identical staining conditions. The slices were incubated with
3% H2O2(5 min),horse sera (10 min) and antibodies (1:750 dilu-
tion for anti-HERV-K antibodies and 1:100 dilution for NCL-5D3
anticytokeratin antibody) (30 min). This was followed by incuba-
tion with anti-rabbit or anti-mouse IgG HRP conjugate antibody
(DAKO) (15 min), incubation with diaminobenzidine (5 min) for
color development and counterstaining with hematoxylin (5 min).
ELISA assays were used to detect anti-HERV antibodies in
human sera as described previously.42Briefly, a 96-well ELISA
plate was coated with various HERV fusion proteins (10 lg/ml,
100 ll/well) including HERV-K Type 1 surface env protein,
HERV-E surface env protein, HERV-K gag protein, ERV3 env
protein and HERV-K plus surface protein, and incubated overnight
at 4?C. After the plate was blocked, human sera (1:200 dilution)
were added and incubated overnight at 4?C. Following 6 washes,
HRP-conjugated anti-human IgG antibody (1:2,000 dilution, Sigma,
St. Louis, MO) was added and color was developed using ABTS
(Sigma) measured on a microplate reader at 405 nm. All ELISA
Expression of multiple HERV env transcripts in human
ovarian cancer cells and tissues
The expression of env transcripts of Type 1 (1,104 bp) and Type
2 (1,194 bp) HERV-K surface domains was detected in ovarian
cancer cell lines (PA1, SKOV3, OVCA 433, OVCAR3, DOV 13
and OVCA 420), but not in normal ovarian epithelial cells (NOE
113, 114, 116 and 119) or immortalized normal ovarian cell lines
(T29, T72 and T80). An example of the RT-PCR results is de-
picted in Figure 1a. The expression of multiple HERV families,
including ERV3, HERV-E and HERV-K Types 1 and 2, was de-
tected in ovarian cancer tissues at a higher frequency than in unin-
volved ovarian tissues (Fig. 1b). Expression of both Rec (437 bp;
Type 2) and np9 (256 bp; Type 1) RNA was detected in ovarian
cancer cells (DOV13 and SKOV3) and in 7 ovarian serous carci-
noma specimens (Fig. 1c). Both Rec and np9 cDNAs obtained
from DOV13 cells were sequenced and the results are shown in
Figure 1d. These results provide evidence that both types of
WANG-JOHANNING ET AL.
HERV-K mRNA, as well as multiple HERV family mRNAs, are
transcribed in ovarian cancer cell lines and tissues.
Furthermore, 254 ovarian tissue RNAs isolated from various
ovarian specimens were quantified for the expression of HERV-K
env transcripts by real-time RT-PCR. The results of real-time RT-
PCR analyses of these samples are presented in Figure 1e. HERV-
K env expression was significantly greater in tissues from epithelial
tumors including serous, endometrioid, mucinous, clear cell, transi-
tional, squamous and mixed malignant mullerian tumor without
metastasis (p 5 0.012; n 5 121), and greater but not statistically
significant in tissues from malignant epithelial tumor with metasta-
sis (p 5 0.058; n 5 46), relative to expression in normal and benign
ovarian tissues (n 5 19), using a t-test.
Surface expression of HERV-K env protein on ovarian
cancer cells lines
Both cell surface (nonpermeabilized; No-perm) and cytoplas-
mic (permeabilized; Perm) expression of HERV-K env protein
FIGURE 1 – Expression of HERV env mRNAs in ovarian cell lines and tissues by RT-PCR. (a) Cell lines: from left to right, each set of lanes
for a given amplified gene represents the RT-PCR expression pattern using ERV3, HERV-E, HERV-K Type 1 (HERV-K(1)), HERV-K Type 2
(HERV-K(2)) and b-actin primers in OVCAR3 ovarian cancer cells (Lane 1), NOE 114 normal ovarian epithelial cells (Lane 2) and SKOV3
ovarian cancer cells (Lane 3). The final lane in each set is a no-template control (Lane 4). (b) Expression of HERV env mRNA in matched tu-
mor/normal tissues: expression of various HERV env mRNAs was evaluated in 2 cancer tissues (Lanes 1 and 3) with their matched uninvolved
normal ovarian tissues (Lanes 2 and 4) obtained from the same patients. The final lane in each set is a no-template control (Lane 5). (c) Expres-
sion of spliced HERV-K transcripts: expression of Rec or np9 spliced env HERV-K transcripts was detected in ovarian carcinoma tissues. Lanes
1–7: each lane represents an ovarian serous carcinoma specimen. The final lane is a no-template control (Lane 8). (d) Alignment of nucleotide
and amino acid sequences of Rec and np9 from DOV13 ovarian cancer cells. (e) Quantitation of HERV-K env mRNA in various ovarian tissues:
the amount of HERV-K in unknown samples was quantitated using CTvalues of HERV-K env mRNA obtained from each specimen, normalized
on the basis of the CTof Homo sapiens ribosomal protein S9. The ratio of HERV-K mRNA CTin ovarian tumor tissues to the average CTin nor-
mal ovarian control tissues was calculated. From left to right: (1) benign epithelial tumor, n 5 6; (2) epithelial tumor with low potential, n 5 5;
(3) mixed epithelial tumor, n 5 11; (4) normal ovarian controls, n 519; (5) placenta (as a control), n 5 13; (6) uninvolved ovarian tissues, n 5
12; (7) epithelial tumor with metastasis, n 5 46; (8) epithelial tumor without metastasis, n 5 121 and (9) sex cord and stromal tumor (including
germ cell tumor), n 5 21. HERV-K env expression was significantly greater (lower CT) in tissues from epithelial tumor without metastasis (p 5
0.012), and epithelial tumor with metastasis (p 5 0.058), relative to expression in normal and benign ovarian control tissues.
EXPRESSION OF HERV env GENES IN OVARIAN CANCER
was detected in OVCAR3 and DOV13 ovarian surface epithelial
cancer cells, but not in T29 and T80 normal cells (Fig. 2a). Flow
cytometric analysis revealed greater surface expression of HERV-
K env protein on nonpermeabilized OVCAR3 (62.22%) and
DOV13 ovarian cancer cells (39.16%), than on nonpermeabilized
T29 and T80 immortalized normal ovarian cells (4–8%; Fig. 2a).
Intracellular expression of HERV-K env surface protein
in ovarian tumor epithelial cells
We next confirmed HERV-K env protein expression in ovarian
cancer tissues, and determined its association with disease stage.
Immunohistochemistry using anti-HERV-K env protein antibody
was performed on multiple tissue microarrays (TMAs). TMA1,
TMA2 and TMA3 contain 72, 85 and 484 multiple ovarian tissues,
respectively. A score of ‘‘0’’ indicates no expression, ‘‘1’’ indicates
low expression and ‘‘2 1 3’’ indicates intermediate and strong
expression of HERV-K env protein, respectively. Examples of
positively stained samples in TMA1 are shown in Figure 2b, while
HERV-K positive samples from TMA2 are shown in Figure 2c.
Expression of HERV-K env SU protein increased in a stepwise
fashion from grade I (33%) to grade II (38%) to high-grade (47%)
serous papillary adenocarcinoma (Fig. 3a) for 40 serous papillary
adenocarcinoma tissues obtained from the TMA1 microarray.
Microarray TMA2 was used for analysis of progression of ovarian
cancer (Fig. 3b). LMP serous, LG serous and LG endometrial
tumors showed higher levels of expression compared to normal
ovaries (p < 0.001). HG serous and endometrial tumors showed
great variability in protein expression with a median expression
slightly lower than normal ovaries. Furthermore, TMA3, contain-
ing 484 cases of various ovarian cancer tissues with clinical fol-
low-up information, was used to assess whether activation of
HERV-K env surface protein correlated with clinical or histologi-
cal characteristics, or with prognostic factors associated with the
patients. Tumor histotype correlated with a significant increase in
expression of HERV-K. A summary of TMA 1 and TMA3 results
is presented in Table I.
Detection of anti-HERV antibodies in sera of ovarian
ELISA analysis was used to detect anti-HERV antibodies in
sera obtained from ovarian cancer patients (n 5 60) and normal
female controls (n 5 50). Approximately 55% of the 60 ovarian
cancer patient samples had positive titers for antibodies against
HERV-K surface protein, 40% were positive for antibodies against
HERV-E surface protein and 30% were positive for antibodies
against ERV3 env protein. The frequency of anti-HERV antibod-
FIGURE 1 – CONTINUED.
WANG-JOHANNING ET AL.
ies from 20 sera obtained from patients with ovarian cancer and
20 normal female controls is summarized in Table II. The pres-
ence of anti-HERV env protein antibodies provides indirect evi-
dence of the presence of HERV proteins in human ovarian cancer.
In the current study, we detected the expression of both types of
HERV-K env proteins in the majority of ovarian cancer cell lines
and tissues analyzed. Sequence analysis revealed that the HERV-K
env cDNAs obtained from 2 serous adenocarcinoma ovarian cancer
specimens contained ORFs and no (Type 1; data not shown) or 1
(Type 2) stop codon. Furthermore, both HERV-K env spliced np9
and Rec mRNA were expressed in ovarian cancer tissues and the
ovarian cancer cell lines DOV13 and OVCAR 3. Rec (Type 2) pro-
tein is localized to the nucleus, and possesses functions similar to
the Rev protein of HIV.35,43Rec was reported in earlier studies to
support cell transformation and to promote tumor formation in
nude mice.34Mice that inducibly express the Rec have altered germ
cell development that may promote carcinoma in situ.44The poten-
tial impact of the unique splice variants that we have described on
ovarian tumorigenesis will require further investigation.
We also observed that other HERV classes (HERV-E and ERV3)
were simultaneously expressed in the same ovarian cancer tissues
that were positive for the expression of HERV-K. It is of interest
that multiple HERV families are transcriptionally active in ovarian
cancer, but to our knowledge not in breast and prostate cancers.
The basis of the preferential transcription of Class I (HERV-E and
ERV3) and Class II (HERV-K) env genes in ovarian cancer tissues
that we describe here is not apparent. However, the selective tran-
scription of HERV-K, HERV-E and ERV3 could be explained by
changes in tumor-specific transcription factors or tissue specific
FIGURE 2 – Expression of HERV env proteins in various ovarian cells and tissues. (a) Localization of HERV-K env protein expression in
ovarian cell lines: Surface and cytoplasmic expression of HERV-K env protein was detected in ovarian cancer cell lines OVCAR3 and DOV13,
without permeabilization (No-perm) or permeabilized with 0.1% Triton X-100 (Perm). IgG FITC antibody only was used as a negative control.
Surface expression of HERV-K env protein was detected in OVCAR3 and DOV13 cells by FACS analysis. Gray areas represent anti-IgG FITC
antibody without anti-HERV-K antibody, as negative controls. Expression of HERV-K env protein was not detected in T80 and T29 normal
ovarian epithelial cells, without permeabilization (No-perm) or permeabilized with 0.1% Triton X-100 (Perm). Surface expression of HERV-K
env protein in was detected in T80 and T29 cells by FACS analysis. (b) Expression of HERV env proteins in various ovarian tissues: scoring of
HERV-K immunostaining from TMA1 microarray: (1) normal ovarian tissues (score ‘‘0’’; 340); (2) clear cell carcinoma (score ‘‘1’’); (3) se-
rous papillary cystadenocarcinoma (score ‘‘2’’) and (4) serous papillary adenocarcinoma (score ‘‘3’’). (c) Samples exhibiting positive immuno-
staining for HERV-K from TMA2 microarray: (1) mucinous cyst; (2) mucinous LMP (low malignant potential); (3) LG (low-grade) endome-
trioid; (4) HG (high-grade) endometrioid; (5) serous LMP; (6) LG serous; (7) HG Serous and (8) clear cell carcinoma.
EXPRESSION OF HERV env GENES IN OVARIAN CANCER
factors, as has been reported for ERV345and HERV-K Group II.28
Activation may also be attributable to selective hypomethylation of
these genes. Altered hypomethylation of HERV-K has been de-
scribed in urothelial carcinoma46and breast cancer (unpublished
data by our group), and HERV-W is hypomethylated in human
ovarian cancer.47Additional studies of a larger number of malignant
and nonmalignant human tissues are needed to determine the mech-
anism, whereby HERV-K, HERV-E and ERV3 are selectively ex-
pressed in ovarian cancer.
HERV gene expression was reported in normal ovarian tissue
by Seifarth et al.48We did not detect HERV env gene expression
in normal ovarian biopsies. However, low levels of expression of
HERV-E or HERV-K env mRNAs in matched uninvolved normal
ovarian tissues was detected by RT-PCR. The positive signal de-
tected in some of these samples by RT-PCR was likely attributable
to the presence of previously undetected malignant cells. Sequence
analysis detected multiple stop codons in uninvolved ovarian tis-
sues, which indicates that no full length env protein can be trans-
lated. In contrast, stop codons were not observed in env mRNAs
isolated from ovarian cancer tissues. These results are further con-
firmed by immunohistochemistry. Env protein expression was not
detected in uninvolved normal ovarian epithelial cells, but clear
expression was observed in tumor epithelial cells in the same biop-
sies. Our transcriptional studies and those of Seifarth et al. showed
low mRNA expression of HERV-K in normal ovarian tissues.
However, expression of HERV-K was not observed at the level of
translation in normal ovarian tissues.
In the Seifarth et al. study, most active members of HERV fam-
ilies were found to be expressed in skin, placenta and tissue of
reproductive organs.48Of 19 human tissues investigated none
could be found that lacked HERV transcription. However, a few
HERV elements have been reported that possess intact ORFs and
the capability to encode functional proteins with potentially patho-
genic activities. For example, expression of spliced env and Rec
mRNA has thus far been observed only in teratocarcinoma and
melanoma tumors. Transcripts of subgenomic HERV-K env were
also detected in the breast cancer cell line T47D and in some
breast cancer tissues. Spliced env and Rec mRNAs are required to
express the env protein, thus explaining why the expression of
HERV-K env protein is detected in only a minority of cases. The
presence of anti-HERV antibodies in the sera of ovarian cancer
patients but not control groups also provides indirect evidence that
HERV env proteins are expressed only in cancer patients, but not
in normal controls.
Recently, expression of HERV-K env protein was demonstrated
at the cell surface of env-transduced cells using anti-HERV-K
FIGURE 2 – CONTINUED.
WANG-JOHANNING ET AL.
antibody.6In an earlier study, when HERV-K envelope protein
ORFs were expressed from a human cytomegalovirus promoter-
driven expression vector in HeLa cells after transient transfection,
expression was observed on both the cell surface and in the cyto-
plasm.6However, no signal was detected when these cells were
transfected with an irrelevant expression vector, which suggests
that high levels of transcript induce a signal for expression both
cytoplasmically and at the cell surface, and argues against a trans-
formation-associated mechanism. Future studies might reveal a
mechanism, whereby transformation of some cell types leads to
induction of cell surface expression. We report here for the first
time that HERV-K env protein is expressed on the surface and in
the cytoplasm of HERV-K-positive ovarian cancer cells not trans-
duced with an HERV-K expression vector (Fig. 2a). Importantly,
no expression was detected in normal or immortalized normal sur-
face ovarian epithelial cells. The expression of surface env protein
was enhanced in ovarian cancer cells after the cells were stimu-
lated with hormone (data not shown). In addition, HERV-K, and
HERV-E envelope, SU protein was detected in the same tumor
epithelial cells that were confirmed by NCL-5D3 monoclonal anti-
body to express cytokeratins 8 and 18 (data not shown). This indi-
cates that functional HERV-K or E env proteins are present on the
ovarian cancer cell surface. This cell surface expression is a pre-
requisite for infectious retrovirus production or generation of an
Multiple TMA slides provided us with a means to compare the
expression of HERV-K env SU protein in multiple tissues under id-
entical conditions of staining. The increase in expression of HERV-
K env SU protein took place in a stepwise fashion in serous papil-
lary adenocarcinoma (Fig. 3a). Expression of HERV-K env SU pro-
tein was significantly increased in LMP serous tumors and endome-
triod tumors, compared with normal ovaries (Fig. 3b). In addition, a
significant increase in expression of HERV-K was correlated with
tumor histotype, especially for epithelial tumors (Table I).
Our studies demonstrated the presence of antibodies against
HERV-K env protein, HERV-E env protein or ERV3 env protein
in the serum of patients with ovarian cancer, but not in normal con-
trols. The generation of HERV-specific antibodies indicates a lack
of tolerance and suggests that expression during ontogenesis does
not occur for ovarian cancer, as suggested by Buscher for HERV-K
expression in melanoma.49Our results show that cancer patients
are capable of producing anti-HERV antibodies, and the presence
of these antibodies indicates that ovarian cancer patients are able to
mount an immune response against specific HERVs, which points
to the potential of immunotherapy for ovarian cancer with HERV
proteins. The role of antibodies against HERV proteins produced
by HERV-positive cancer patients needs to be further investigated.
The widespread distribution of HERVs in mammalian genomes
implies that they may influence host pathophysiology. HERVs
could be envisioned to contribute to pathogenicity as a partner in
recombination events in several ways: by de novo insertion after
mobilization, followed by activation of downstream proto-onco-
genes or by gene disruption. De novo insertion of HERV genes or
proviruses has not been shown up to now in single somatic cells
(e.g. tumors) and it may be a rare event. A possibly more relevant
mechanism of ‘‘endogenous retroviral oncogenesis’’ than proviral
insertion (or promotor insertion) could be the recombination of
HERV env genes to an oncogenic entity, as has been shown in the
spontaneous development of highly leucemogenic MCV viruses in
AKR mice by recombination of different MuLV env genes.50Con-
sistent with this possibility is the observation that the coding
sequence of the apoptosis-controlling gene BNIP3 is made up
almost entirely of sequences from an apparent human endogenous
retrovirus, suggesting that fragments of mobile elements have been
inserted into functioning genes.51
Reexpression of viral proteins accompanied by loss of immune
tolerance could induce immune disturbances.52This is the first
report of expression of multiple HERV env proteins in ovarian can-
cer tissues. In our study, we have demonstrated that HERV is ex-
pressed in ovarian cancer cells and tissues at both the mRNA and
protein levels. Likewise, we are the first to quantitate expression of
the HERV env protein in ovarian cancer cells and to assess expres-
sion in a large number of tissue samples grouped together for si-
multaneous uniform staining. The most promising findings of our
study are that multiple types of HERV are expressed simultane-
ously, that antibodies to HERVs are present in the sera of ovarian
cancer patients and that HERVs are expressed on the surface of
ovarian cancer cells. The presence of serum antibodies suggests
that HERV proteins may be immunogenic and therefore capable of
acting as tumor-associated antigens.
Activation of HERV-K expression in ovarian cancer might occur
in response to a transcription factor found specifically in malignant
ovarian epithelial cells that promotes LTR-directed transcription of
HERV-K during transformation and tumor progression.53–55Sec-
ond, such activation might occur as a result of hypomethylation of
HERV genomic DNA during transformation and tumor progression.
Retrotransposons have been identified as potential targets of hypo-
methylation during cellular transformation.47Enhanced HERV-K
expression has been reported to result from DNA hypomethylation
in urothelial carcinomas46and germ cell tumors.56A similar mecha-
nism may occur in ovarian carcinoma.
Given the high frequency of expression (90%) of the HERV-K
env protein in ovarian epithelial tumors and the lack of expression
FIGURE 3 – Expression profile of HERV-K env protein in ovarian
tissues. (a) HERV-K env SU protein expression in serous papillary ad-
enocarcinoma of various grades (I, II and III). Percentage of ‘‘no
expression’’ progressively decreased from lower to higher grades,
whereas percentage of ‘‘strong expression’’ progressively increased
from lower to higher grades. (b) Analysis of ovarian cancer progres-
sion with TMAs: (1) normal ovary; (2) mucinous cyst; (3) mucinous
tumor of LMP; (4) serous tumor of LMP; (5) low-grade serous carci-
noma; (6) low-grade endometrial carcinoma; (7) high-grade serous
carcinoma; (8) high-grade endometrial carcinoma and (9) clear cell
carcinoma. LMP and low-grade tumors showed higher levels of
expression compared to normal ovarian surface epithelial cells (Krus-
kall Wallis analysis p < 0.001). High-grade tumors showed great vari-
ability in protein expression with a median expression slightly lower
compared to normal ovaries.
EXPRESSION OF HERV env GENES IN OVARIAN CANCER
in normal and benign ovarian surface epithelial tissues in the pres-
ent study, HERV-K has the potential to be an important ovarian
cancer-associated antigen that could be used for immunotherapy
against ovarian cancer. Evaluation of HERV expression may thus
provide a new ovarian cancer screening tool, and serve as a novel
target for detection, diagnosis, and treatment of ovarian cancer.
We thank Drs. Robert C. Bast Jr. and Yinhua Yu (M.D.
Anderson Cancer Center) for providing us with ovarian cell
lines. We thank the Cattleman for Cancer Research Foundation
for providing us with a fluorescence microscope.
TABLE II – THE FREQUENCY OF ANTI-HERV ANTIBODIES IN OVARIAN CANCER PATIENTS OR NORMAL FEMALE CONTROLS
Age/ethnic group/sexDiagnosis K-SUE-SU K-gag K-Plus ERV3
Ovarian cancer patients summary (N 5 20)
Clear cell carcinoma
Papillary serous carcinoma
Recurrent ovarian carcinoma
Papillary serous carcinoma
Recurrent ovarian carcinoma
Mestatic serous papillary carcinoma
Papillary serous carcinoma
Metastatic serous papillary carcinoma
Metastatic serous papillary carcinoma
OD > 1
OD 5 0.75–1
OD 5 0.5–0.75
OD < 0.5
Total positive %
Normal female controls summary (N 5 20)
OD > 1
OD 5 0.75–1
OD 5 0.5–0.75
OD < 0.5
Total positive %
K-SU, HERV-K (type 1) env surface protein; E-SU, HERV-E env surface protein; K-gag, HERV-K gag protein; K-Plus, HERV-K (type 2) env
surface protein with 292 bp insert; ERV3, ERV3 env protein; 2, no antibody presence (OD < 0.5); 1, low antibody presence (OD 5 0.5–0.75);
11, intermediate antibody presence (OD 5 0.75–1); 111, high antibody presence (OD > 1).
TABLE I – THE EXPRESSION PROFILE OF HERV-K ENV SU PROTEIN IN OVARIAN TISSUE MICROARRAY SLIDES
0*12 1 3
Histotype TMA1 (N 5 72)
Normal ovarian tissues
Germ cell tumors
Granular cell tumors
Mucous papillary adenocarcinoma
Serous P AdCa
Clear cell carcinoma
Histotype TMA3 (N 5 484)
Malignant mixed m€ ullerian tumor
Clear cell carcinoma
Poorly differentiated carcinoma
Transitional cell carcinoma
1 (20%)1 (20%)
*A score of 0 indicates no expression, 1 indicates low expression and 213 indicates intermediate and
strong expression.–**p values were calculated by using a v2test of independence.
WANG-JOHANNING ET AL.
1. Lander ES, Linton LM, Birren B, Nusbaum C, Zody MC, Baldwin J,
Devon K, Dewar K, Doyle M, FitzHugh W, Funke R, Gage D et al.
Initial sequencing and analysis of the human genome. Nature 2001;
Hughes JF, Coffin JM. Evidence for genomic rearrangements medi-
ated by human endogenous retroviruses during primate evolution. Nat
Griffiths DJ. Endogenous retroviruses in the human genome sequence.
Genome Biol 2001;2:REVIEWS1017.
Andersson AC, Yun Z, Sperber GO, Larsson E, Blomberg J. ERV3 and
related sequences in humans: structure and RNA expression. J Virol
Bessis D, Moles JP, Basset-Seguin N, Tesniere A, Arpin C, Guilhou JJ.
Differential expression of a human endogenous retrovirus E transmem-
brane envelope glycoprotein in normal, psoriatic and atopic dermatitis
human skin. Br J Dermatol 2004;151:737–45.
Dewannieux M, Blaise S, Heidmann T. Identification of a functional
envelope protein from the HERV-K family of human endogenous ret-
roviruses. J Virol 2005;79:15573–7.
Blaise S, de Parseval N, Benit L, Heidmann T. Genomewide screening
for fusogenic human endogenous retrovirus envelopes identifies syncy-
tin 2, a gene conserved on primate evolution. Proc Natl Acad Sci USA
Blaise S, Ruggieri A, Dewannieux M, Cosset FL, Heidmann T. Identi-
fication of an envelope protein from the FRD family of human endog-
enous retroviruses (HERV-FRD) conferring infectivity and functional
conservation among simians. J Virol 2004;78:1050–4.
Jern P, Lindeskog M, Karlsson D, Blomberg J. Full-length HERV-H
elements with env SU open reading frames in the human genome.
AIDS Res Hum Retroviruses 2002;18:671–6.
10. de Parseval N, Lazar V, Casella JF, Benit L, Heidmann T. Survey of
human genes of retroviral origin: identification and transcriptome of
the genes with coding capacity for complete envelope proteins. J Virol
11. O’Connell C, O’Brien S, Nash WG, Cohen M. ERV3, a full-length
human endogenous provirus: chromosomal localization and evolution-
ary relationships. Virology 1984;138:225–35.
12. O’Connell CD, Cohen M. The long terminal repeat sequences of a
novel human endogenous retrovirus. Science 1984;226:1204–6.
13. Cohen M, Powers M, O’Connell C, Kato N. The nucleotide sequence
of the env gene from the human provirus ERV3 and isolation and char-
acterization of an ERV3-specific cDNA. Virology 1985;147:449–58.
14. de Parseval N, Heidmann T. Physiological knockout of the envelope
gene of the single-copy ERV-3 human endogenous retrovirus in a
fraction of the Caucasian population. J Virol 1998;72:3442–5.
15. Andersson AC, Merza M, Venables P, Ponten F, Sundstrom J, Cohen M,
Larsson E. Elevated levels of the endogenous retrovirus ERV3 in human
sebaceous glands. J Invest Dermatol 1996;106:125–8.
16. Boyd MT, Bax CM, Bax BE, Bloxam DL, Weiss RA. The human en-
dogenous retrovirus ERV-3 is upregulated in differentiating placental
trophoblast cells. Virology 1993;196:905–9.
17. Li JM, Fan WS, Horsfall AC, Anderson AC, Rigby S, Larsson E,
Venables PJ. The expression of human endogenous retrovirus-3 in fe-
tal cardiac tissue and antibodies in congenital heart block. Clin Exp
18. Larsson E, Andersson AC, Nilsson BO. Expression of an endogenous
retrovirus (ERV3 HERV-R) in human reproductive and embryonic
tissues—evidence for a function for envelope gene products. Ups J Med
19. Katsumata K, Ikeda H, Sato M, Ishizu A, Kawarada Y, Kato H,
zWakisaka A, Koike T, Yoshiki T. Cytokine regulation of env gene
expression of human endogenous retrovirus-R in human vascular en-
dothelial cells. Clin Immunol 1999;93:75–80.
20. Kitamura M, Maruyama N, Shirasawa T, Nagasawa R, Watanabe K,
Tateno M, Yoshiki T. Expression of an endogenous retroviral gene
product in human placenta. Int J Cancer 1994;58:836–40.
21. Turbeville MA, Rhodes JC, Hyams DM, Distler CM, Steele PE.
Expression of a putative immunosuppressive protein in human tumors
and tissues. Pathobiology 1996;64:233–8.
22. Aguinaldo AM, Turbeville JM, Linford LS, Rivera MC, Garey JR,
Raff RA, Lake JA. Evidence for a clade of nematodes, arthropods and
other moulting animals. Nature 1997;387:489–93.
23. Lower R, Boller K, Hasenmaier B, Korbmacher C, Muller-Lantzsch
N, Lower J, Kurth R. Identification of human endogenous retroviruses
with complex mRNA expression and particle formation. Proc Natl
Acad Sci USA 1993;90:4480–4.
24. Lower R, Lower J, Kurth R. The viruses in all of us: characteristics
and biological significance of human endogenous retrovirus sequen-
ces. Proc Natl Acad Sci USA 1996;93:5177–84.
25. Lower R, Tonjes RR, Korbmacher C, Kurth R, Lower J. Identification
of a Rev-related protein by analysis of spliced transcripts of the
human endogenous retroviruses HTDV/HERV-K. J Virol 1995;69:
26. Seifarth W, Skladny H, Krieg-Schneider F, Reichert A, Hehlmann R,
Leib-Mosch C. Retrovirus-like particles released from the human
breast cancer cell line T47-D display type B- and C-related endoge-
nous retroviral sequences. J Virol 1995;69:6408–16.
27. Etkind PR, Lumb K, Du J, Racevskis J. Type 1 HERV-K genome is
spliced into subgenomic transcripts in the human breast tumor cell
line T47D. Virology 1997;234:304–8.
28. Willer A, Saussele S, Gimbel W, Seifarth W, Kister P, Leib-Mosch C,
Hehlmann R. Two groups of endogenous MMTV related retroviral env
transcripts expressed in human tissues. Virus Genes 1997;15:123–33.
29. Wang-Johanning F, Frost AR, Johanning GL, Khazaeli MB, LoBuglio
AF, Shaw DR, Strong TV. Expression of human endogenous retrovirus
k envelope transcripts in human breast cancer. Clin Cancer Res 2001;
30. Ono M, Kawakami M, Ushikubo H. Stimulation of expression of the
human endogenous retrovirus genome by female steroid hormones in
human breast cancer cell line T47D. J Virol 1987;61:2059–62.
31. Magin C, Hesse J, Lower J, Lower R. Corf, the Rev/Rex homologue
of HTDV/HERV-K, encodes an arginine-rich nuclear localization sig-
nal that exerts a trans-dominant phenotype when mutated. Virology
32. Magin C, Lower R, Lower J. cORF and RcRE, the Rev/Rex and RRE/
RxRE homologues of the human endogenous retrovirus family
HTDV/HERV-K. J Virol 1999;73:9496–507.
33. Yang J, Bogerd HP, Peng S, Wiegand H, Truant R, Cullen BR. An an-
cient family of human endogenous retroviruses encodes a functional
homolog of the HIV-1 Rev protein. Proc Natl Acad Sci USA 1999;
34. Boese A, Sauter M, Galli U, Best B, Herbst H, Mayer J, Kremmer E,
Roemer K, Mueller-Lantzsch N. Human endogenous retrovirus pro-
tein cORF supports cell transformation and associates with the promy-
elocytic leukemia zinc finger protein. Oncogene 2000;19:4328–36.
35. Boese A, Sauter M, Mueller-Lantzsch N. A Rev-like NES mediates
cytoplasmic localization of HERV-K cORF. FEBS Lett 2000;468:65–7.
36. Wang-Johanning F. Quantitation of HERV-K env gene expression
and splicing in human breast cancer. Oncogene 2003;22:1528–35.
37. Wang-Johanning F, Frost AR, Jian B, Azerou R, Lu DW, Chen DT,
Johanning GL. Detecting the expression of human endogenous retrovi-
rus E envelope transcripts in human prostate adenocarcinoma. Cancer
38. Liu J, Yang G, Thompson-Lanza JA, Glassman A, Hayes K, Patterson A,
Marquez RT, Auersperg N, Yu Y, Hahn WC, Mills GB, Bast RC, Jr. A
genetically defined model for human ovarian cancer. Cancer Res 2004;
39. Mayer J, Sauter M, Racz A, Scherer D, Mueller-Lantzsch N, Meese E.
An almost-intact human endogenous retrovirus K on human chromo-
some 7. Nat Genet 1999;21:257–8.
40. Armbruester V, Sauter M, Krautkraemer E, Meese E, Kleiman A,
Best B, Roemer K, Mueller-Lantzsch N. A novel gene from the
human endogenous retrovirus K expressed in transformed cells. Clin
Cancer Res 2002;8:1800–7.
41. Wang-Johanning F, Lu DW, Wang Y, Johnson MR, Johanning GL.
Quantitation of human papillomavirus 16 E6 and E7 DNA and RNA
in residual material from ThinPrep Papanicolaou tests using real-time
polymerase chain reaction analysis. Cancer 2002;94:2199–210.
42. Wang-Johanning F, Gillespie GY, Grim J, Rancourt C, Alvarez RD,
Siegal GP, Curiel DT. Intracellular expression of a single-chain anti-
body directed against human papillomavirus type 16 E7 oncoprotein
achieves targeted antineoplastic effects. Cancer Res 1998;58:1893–900.
43. Tonjes RR, Lower R, Boller K, Denner J, Hasenmaier B, Kirsch H,
Konig H, Korbmacher C, Limbach C, Lugert R, Phelps RC, Scherer J
et al. HERV-K: the biologically most active human endogenous retro-
virus family. J Acquir Immune Defic Syndr Hum Retrovirol 1996;
44. Galli UM, Sauter M, Lecher B, Maurer S, Herbst H, Roemer K,
Mueller-Lantzsch N. Human endogenous retrovirus rec interferes
with germ cell development in mice and may cause carcinoma in situ,
the predecessor lesion of germ cell tumors. Oncogene 2005;24:
45. Kato N, Pfeifer-Ohlsson S, Kato M, Larsson E, Rydnert J, Ohlsson R,
Cohen M. Tissue-specific expression of human provirus ERV3 mRNA
in human placenta: two of the three ERV3 mRNAs contain human
cellular sequences. J Virol 1987;61:2182–91.
46. Florl AR, Lower R, Schmitz-Drager BJ, Schulz WA. DNA methyla-
tion and expression of LINE-1 and HERV-K provirus sequences in
urothelial and renal cell carcinomas. Br J Cancer 1999;80:1312–21.
47. Menendez L, Benigno BB, McDonald JF. L1 and HERV-W retrotrans-
posons are hypomethylated in human ovarian carcinomas. Mol Cancer
EXPRESSION OF HERV env GENES IN OVARIAN CANCER
48. Seifarth W, Frank O, Zeilfelder U, Spiess B, Greenwood AD,
Hehlmann R, Leib-Mosch C. Comprehensive analysis of human en-
dogenous retrovirus transcriptional activity in human tissues with a
retrovirus-specific microarray. J Virol 2005;79:341–52.
49. Buscher K, Trefzer U, Hofmann M, Sterry W, Kurth R, Denner J.
Expression of human endogenous retrovirus K in melanomas and mel-
anoma cell lines. Cancer Res 2005;65:4172–80.
50. Holland CA, Hartley JW, Rowe WP, Hopkins N. At least four viral
genes contribute to the leukemogenicity of murine retrovirus MCF
247 in AKR mice. J Virol 1985;53:158–65.
51. Britten RJ. Coding sequences of functioning human genes derived
entirely from mobile element sequences. Proc Natl Acad Sci USA
52. Lower R. The pathogenic potential of endogenous retroviruses: facts
and fantasies. Trends Microbiol 1999;7:350–6.
53. de Parseval N, Alkabbani H, Heidmann T. The long terminal repeats
of the HERV-H human endogenous retrovirus contain binding sites
for transcriptional regulation by the Myb protein. J Gen Virol 1999;
54. Lee WJ, Kwun HJ, Kim HS, Jang KL. Activation of the human en-
dogenous retrovirus W long terminal repeat by herpes simplex virus
type 1 immediate early protein 1. Mol Cells 2003;15:75–80.
55. Sjottem E, Anderssen S, Johansen T. The promoter activity of long
terminal repeats of the HERV-H family of human retrovirus-like ele-
ments is critically dependent on Sp1 family proteins interacting with a
GC/GT box located immediately 30to the TATA box. J Virol 1996;
56. Lavie L, Kitova M, Maldener E, Meese E, Mayer J. CpG methylation
directly regulates transcriptional activity of the human endogenous
retrovirus family HERV-K(HML-2). J Virol 2005;79:876–83.
WANG-JOHANNING ET AL.