The opioid growth factor (OGF) and its receptor (OGFr) regulate human ovarian cancer cell proliferation through a tonically active inhibitory axis. We investigated the effect of OGFr overexpression on ovarian tumorigenesis. Clonal cell lines of SKOV-3 human ovarian cancer were established to stably overexpress OGFr. shRNA constructs were evaluated for antitumor activity in vitro, as well as in vivo using mouse models of s.c. and i.p. tumor transplantation. The 5 clonal cell lines were characterized by increases in OGFr protein (62% to 245%) and binding capacity (51% - 154%), and decreases (36% - 185%) in cell number, relative to untransfected wild-type (WT) cells and empty vector (EV) transfected clones. Nude mice receiving s.c. injection of 2 overexpressing OGFr cell lines (OGFr-3 and OGFr-22) had reduced tumor incidence, delayed tumor appearance (up to 12 days), and decreased tumor volume (up to 87%) relative to WT and EV controls. Mice injected i.p. with these clonal lines displayed reduced formation of tumor nodules (up to 95%), and depressed tumor weights (up to 99%) compared to WT and EV groups. DNA synthesis, but not cell survival, was depressed in cells and s.c. tumors overexpressing OGFr in comparison to the WT and EV groups. Angiogenesis was reduced up to 86% in clonal tumors compared to WT and EV groups. This preclinical evidence demonstrates that OGFr expression is a molecular determinant of ovarian cancer progression, and has important relevance to understanding the pathogenesis and treatment of this deadly disease.
Ovarian cancer is the 5th leading cause of cancer related mortality among women in the United States, and the leading cause of death from gynecological malignancies [
An integral component of the ovarian cancer phenotype is dysregulation of cell proliferation [
The relationship of endogenous opioids and ovarian cancer has received some attention. In 1982, Sporrong et al. [
Donahue and collaborators [
The human ovarian cancer cell line, SKOV-3 [
SKOV-3 cells were transfected with pcDNA3.1+ vector (empty vector, EV) or with the plasmid pcDNA3.1+ human OGFr in the presence of lipofectamine 2000 (Invitrogen, Carlsbad, CA) for 4 h in serum and antibiotic free media. At 4 h, cultures were supplemented with serum containing media. At 24 h, transfection reagents were removed and replaced with serum containing media. Transfected cells were selected by growth in media containing G418 at 500 µg/ml; 24 clones were expanded and analyzed by Western blotting. Based on OGFr expression, 5 clones (OGFr-3, OGFr-13, OGFr-15, OGFr-21, and OGFr-22) were maintained and further characterized by semiquantitative immunohistochemistry, Western blot, receptor binding assays, and growth. For all experiments, untransfected wild-type (WT) cells and EV transfected clones served as controls.
Clonal cells, as well as WT and EV cells, were plated and counted 24 h later (time 0) to determine seeding efficiency. For treatment studies, 10–6 M OGF or naltrexone (NTX) (Sigma Aldrich, St. Louis, MO) was added at time 0; media and compounds were replaced daily. Drugs were prepared in sterile water and dilutions represent final concentrations of the compounds. An equivalent volume of sterile water was added to controls. At designated times, cells were harvested, stained with trypan blue, and counted with a hemacytometer. At least two aliquots/well from at least 2 wells/treatment/timepoint were sampled.
Four week-old athymic nu/nu female mice, purchased from Charles River Laboratory (Wilmington, MA), were housed in pathogen-free isolator ventilated cages in a controlled-temperature room (22˚C - 25˚C) with a 12 - 12 h light/dark cycle (lights on 0700 - 1900) in the Department of Comparative Medicine at The Pennsylvania State University College of Medicine. Sterile water and standard rodent diet (Harlan Teklad, Fredrick, MD) were available ad libitum. All procedures were approved by the IACUC committee of The Pennsylvania State University College of Medicine, and conformed to the guidelines established by the NIH. Mice were allowed 48 h to acclimate prior to experimentation.
Clonal cell lines stably overexpressing OGFr, OGFr-3, and OGFr-22, as well as EV and WT cells, were expanded and analyzed by receptor binding to determine the binding capacity of OGFr prior to inoculation into nude mice. For the subcutaneous xenograft model, 4 × 106 cells of four different lines (WT, EV, OGFr-3, or OGFr-22) were injected into the right scapula region of unanaesthetized mice; 12 mice per cell line were used. For the intraperitoneal xenograft model, unanaesthetized mice were injected with 5 × 106 EV, OGFr-3, or OGFr- 22 cells; 8 - 12 mice were inoculated with each cell line. These concentrations were selected based on published reports [31-34] as well as preliminary tumor burden studies (Donahue et al., unpublished observations).
Mice with subcutaneous xenografts were weighed weekly and observed daily for initial appearance of a visible tumor. The latency for a visible tumor and the time until tumors were measurable (≥ 62.5 mm3) were recorded. Tumors were measured in two dimensions with vernier calipers 3 times/week. Volume was calculated using the formula l × w2 × π/6, where length (l) is the longest dimension, and width (w) is the dimension perpendicular to the length [
Mice were euthanized 32 and 40 days following subcutaneous or intraperitoneal tumor cell inoculation, respectively, by an overdose of sodium pentobarbital (100 mg/kg) and cervical dislocation. To examine DNA synthesis in tumors, a subset of mice from each group was injected intraperitoneally with BrdU (100 mg/kg) at 6 and 3 h prior to euthanization. Mice with subcutaneous xenografts were weighed, tumors and spleens were removed and weighed, and the lymph nodes, liver, and spleen were examined for metastases. Tumors were processed for immunohistochemistry, BrdU, hematoxylin and eosin, and TUNEL analysis. Mice with intraperitoneal xenografts were weighed and the number of tumor nodules on the surfaces of the liver, stomach, spleen, and intestines were recorded, removed, and weighed.
Semiquantitative immunohistochemistry was utilized to evaluate the presence and relative levels of OGF and OGFr in cells and tumor tissue according to Donahue et al. [
Expression of OGFr was evaluated in clonal lines by Western blot according to published procedures [
Log phase cells were assayed for OGFr using custom synthesized [3H]-[Met5]-enkephalin (Perkin Elmer-New England Nuclear; 52.7 Ci/mmol) following procedures by Donahue et al. [
The effect of overexpressing OGFr on DNA synthesis, apoptosis, and necrosis was assessed in cells. Tumor tissue was evaluated for DNA synthesis and apoptosis. Cells (5 × 104/coverglass) were grown in culture for 72 h, pulsed with BrdU (30 μM, Sigma Aldrich) for 3 h, and fixed in formalin. Tumors from mice receiving BrdU were fixed in formalin overnight, processed in paraffin, and sectioned at 10 µm. Cells or tissue was stained with anti-BrdU antibody (1:200, Invitrogen) to assess DNA synthesis [11,23,24], or processed for TUNEL according to the manufacturer’s instruction (Trevigen, Gaithersburg, MD) to measure apoptosis [
Staining with hematoxylin and eosin was performed on tumor tissue to examine endothelial lined vessels containing red blood cells [37,38]. Blood vessel density was determined from at least 10 random fields around the periphery of each tumor, with 2 sections/tumor, and 2 tumors/treatment group evaluated.
Tumor incidence was analyzed using the Chi square test; all other data were analyzed using one-way analysis of variance (ANOVA) with subsequent comparisons made using Newman-Keuls tests (GraphPad Prism). In some cases, data were evaluated using unpaired t-tests. P values < 0.05 were considered to be significant.
To study the effects of amplification of OGFr on the growth of ovarian cancer cells, SKOV-3 cells were stably transfected with an OGFr expression vector; 24 neomycin-resistant clones were initially characterized by Western blot (data not shown). Five clones with varying levels of OGFr expression as compared to the WT and EV cultures were expanded and further characterized by semiquantitative immunohistochemistry, Western blot, OGF receptor binding, and growth (Figures 1 and 2). For all studies, comparisons were made to WT and EV groups.
For all cultures, OGFr was visible in the cytoplasm and a speckling of immunoreactivity often noted in cell nuclei (
To further characterize OGFr overexpressing cell lines, binding capacity (Bmax) and binding affinity (Kd) of OGFr for radiolabeled [Met5]-enkephalin was determined. Specific and saturable binding was identified in the nuclear fraction of all cell lines. Bmax values for clonal cells were markedly increased compared to those of WT (3.84 ± 0.21) and EV (3.45 ± 0.29) groups, with increases ranging from 51% - 154% in cells with amplified OGFr (
The functional repercussions of overexpressing OGFr were determined by evaluating cell growth. Cell number was significantly decreased 36% - 85% over a 120 h period in clonal lines overexpressing OGFr compared to EV and WT controls (
In earlier studies the addition of exogenous OGF to ovarian cancer cultures depressed cell number [11,21], leading to the prediction that exogenous OGF introduced to cells with amplification of OGFr would result in an exaggerated inhibitory response. Under standard growth conditions, the basal number of cells in cultures overexpressing OGFr was reduced 70% - 85% compared to WT and EV cultures (
To understand the effects of opioid receptor blockade with a potent and long-acting opioid receptor antagonist on cells with an abundance of OGFr, 10–6 M NTX was added to the cultures (
To evaluate the mechanism by which an excess of OGFr decreases ovarian cancer cell number, DNA synthesis and cell survival were evaluated. In comparison to the BrdU labeling index of WT (26%) and EV (25%) cultures, the labeling index in clonal cell lines overexpressing OGFr was decreased 59% - 68% (
To examine OGF levels in cultures overexpressing OGFr, semiquantitative immunohistochemistry for OGF was performed. OGF was visible in the cytoplasm with a speckling of immunoreactivity often noted in cell nuclei (
Measurable tumors (i.e., >62 mm3) began to form 4 days following tumor cell inoculation in mice receiving WT or EV cells. One day later, when 90% and 100% of mice inoculated with WT and EV cells, respectively, had measurable tumors, 0% of mice inoculated with OGFr-3 or OGFr-22 cells had measurable tumors (
By days 10 and 20, when 100% of mice in the WT and EV groups had measurable tumors, only 10% and 30%, respectively, of mice receiving OGFr-3 cells, and 50% and 80%, respectively, of mice administered OGFr-22 cells, had measurable tumors. By the end of the study (day 32), 40% of mice inoculated with OGFr-3 cells and 90% of mice receiving OGFr-22 cells displayed measurable tumors. An evaluation of the latency to development of measurable tumors revealed latencies of 4.33 ± 0.17 and 4.30 ± 0.15 days for mice administered WT or EV cells, respectively (
Of mice that developed measurable tumors, tumor volumes were decreased 28% - 87% in mice injected with OGFr-3 cells, and reduced 19% - 78% in mice inoculated with OGFr-22 cells, beginning on days 11 and 6, respectively; these measurements persisted throughout the duration of the study compared to WT and EV controls (
Compared to WT and EV controls on the day of termination (day 32), mice inoculated subcutaneously with clones overexpressing OGFr and developing measurable tumors displayed a visible reduction in tumor size (
In the intraperitoneal xenograft model, mice inoculated with clones overexpressing OGFr (OGFr-3 or OGFr-22) displayed a 95% and 65% reduction, respectively, in the total number of tumor nodules compared to EV controls at the end of the 40 day study and ((Figures 5(d) and (e)), reflecting changes in the number of nodules identified on the liver, intestines and stomach. With respect to tumor nodules detected on the spleen, comparable numbers were noted in mice inoculated with OGFr-22 and EV cells; however, there were no nodules the spleen noted in mice inoculated with OGFr-3 cells (
Total tumor weights in mice receiving intraperitoneal injections of OGFr-3 and OGFr-22 cells were reduced 99% and 69%, respectively, relative to EV controls (
Examination of late stage apoptosis in tumors by TUNEL assay revealed similar levels of cell death in mice inoculated with WT, EV, OGFr-3, or OGFr-22 cells (Figures 6(a) and (b)). With respect to DNA synthesis, comparable rates were noted in tumors from mice receiving WT (31.9 ± 1.2) or EV (30.6 ± 1.5) cells. However, DNA synthesis was reduced in tumors from mice in the OGFr- 3 and OGFr-22 groups by 78% and 67%, respectively, compared to WT controls (Figures 6(c) and (d)).
To investigate the distribution and expression of OGFr and OGF in tumors, semiquantitative immunohistochemistry was performed. The location of OGFr was similar in tumors from all groups of mice, with immunoreactivity for this receptor detected in the cytoplasm and a speckling of immunoreactivity noted in cell nuclei (
Relative to tumor vessel density in mice inoculated with WT or EV cells, vessel density was reduced 86% and 65% in mice injected with OGFr-3 or OGFr-22 cells, respectively (Figures 7(e) and (f)).
This study is the first to report the stable molecular overexpression of OGFr in a human ovarian cancer cell line, and reveals that upregulation of OGFr markedly inhibits the proliferation of cells in vitro and tumorigenesis in vivo. In tissue culture, ovarian cancer cells engineered to have an overexpression of OGFr had decreases in cell
number and DNA synthesis, increases in doubling time, and exhibited considerably more modulatory capability in the face of challenges with an opioid agonist (OGF) or antagonist (NTX) compared to WT or EV controls. Xenografts using the subcutaneous route with cells having an abundance of OGFr had increases from control levels in the interval to form a measurable tumor, decreases in the number of animals displaying a measurable tumor, and reductions in tumor volume and weight. With the intraperitoneal route, both the number of metastases and tumor weight were markedly reduced from WT and EV controls. In both in vitro and in vivo investigations, the effects of an excess of OGFr were maintained, indicating that there was no tolerance to the repercussions from amplification of OGFr in these cells.
Evidence of stable overexpression of OGFr in ovarian cancer cells comes from several avenues of experimentation. An increase from control levels (WT, EV) in OGFr protein was detected with semiquantitative immunohistochemistry and Western blotting. The functional capability of the overexpressed OGFr in cellular homogenates to bind to OGF was documented in receptor binding studies, wherein a significant increase from the WT and EV ovarian cancer cells in binding capacity was detected. The overexpression of OGFr in these neoplastic cells did not alter binding affinity of OGF to the receptor, indicating that the processes accompanying the translation of the excess OGFr in these cells was comparable to that in WT cells. Moreover, the binding affinity, as well as binding capacity, in EV cells also was similar to that in WT cells, denoting that the vector did not contribute to any changes in the overexpressed OGFr. Finally, the selection and characterization of multiple clonal cell lines with overexpressed OGFr insured that the outcome of the
transfection was a consistent rather than isolated observation. That the addition of clonal OGFr had a physiological action in cells that was of greater magnitude than in WT cells, suggests that the downstream pathways (e.g., nucleocytoplasmic transport) remained intact and accommodated the excess OGFr.
To investigate the mechanism by which OGFr overexpression inhibits cell number and tumor progression, cell survival and DNA synthesis were assessed. No changes in the number of apoptotic and/or necrotic ovarian cancer cells with molecular amplification of OGFr were discerned either in tissue culture or in xenografts when compared to WT or EV cells/tissues. Therefore, the reduction in tumor size could not be accounted partially or completely by alteration of cell survival pathways. However, DNA synthesis, both under in vitro and in vivo environments, was markedly depressed in the ovarian cancer cells with an abundance of OGFr compared to WT or EV controls. These results are entirely consistent with previous reports showing that the OGF-OGFr axis serves to maintain the pace of cell proliferation through an inhibitory cascade [7-11], with peptide-receptor interaction targeted to upregulating the cyclin dependent inhibitory kinase pathways [11-14].
The OGF-OGFr axis is known to regulate cell proliferative events with respect to the vascular system, including angiogenesis and repair of vascular injury [39- 41]. In the case of overexpression of OGFr in ovarian cancer cells, xenografts were found to have a reduction from control levels in the number of blood vessels associated with these tumors. These data would suggest that the reduction in tumor burden was correlated with a decrease in the vascular supply needed for nutrition. Thus, ovarian tumorigenesis responded to both a direct effect from an excess of OGFr (i.e., a decrease in cell proliferation), and an indirect effect on depressing angiogenesis as a consequence of smaller tumor burden.
As OGF has previously been identified as the opioid that binds to OGFr to inhibit cell proliferation in ovarian cancer [
To evaluate the total magnitude by which the OGFOGFr axis can modulate the growth of ovarian cancer cells, clonal cell lines overexpressing OGFr were treated with exogenous OGF or subjected to continuous opioid receptor blockade with NTX. Cultures exposed to OGF or NTX responded with a decrease or increase, respectively, in cell number compared to cohorts treated with vehicle. Cells overexpressing OGFr, however, displayed a greater than 2-fold enhanced response to opioid receptor antagonism, but a comparable response to exogenous OGF compared to WT or EV controls. If one totals the overall magnitude of response to OGF and NTX, growth regulation by the OGF-OGFr axis was 2.5- to 4.6-fold greater in OGFr overexpressing clones than in WT or EV cultures. Thus, the OGF-OGFr axis has a considerable range of modulatory capability in human ovarian cancer cells.
The present observations on the effects of an abundance of OGFr in ovarian cancer cells with regard to cell proliferation and tumorigenesis complement previous reports on the repercussions of molecular manipulation of OGFr. Earlier studies have shown that transient transfection of OGFr cDNA into rat corneal epithelial cells using a gene gun depressed DNA synthesis [
Clinically, OGF has been detected by radioimmunoassay in surgical samples taken from human neoplasms of the ovary [
The present investigation provides insight into the molecular mechanisms of the OGF-OGFr axis as an integral component controlling ovarian cancer cell proliferation and tumorigenesis, and provides evidence that this system can be exaggerated in ovarian cancer to depress the progression of disease. The findings in this study have clinical importance in designing treatment modalities that take advantage of this biological inhibitory axis.
We are grateful for the support by the Paul K. and Anna E. Shockey Family, Bonnie and Ken Shockey, and the Zagon/Kostel families. The authors declare that they have no competing interests.