By Peter Choong, MD; Crispin Dass, PhD and Matthew L. Broadhead, BSc(Med) MBBS

Osteosarcoma is the most common malignancy of bone and predominately affects young adults. Current treatment regimes aim to control tumour growth and metastasis by combining chemotherapy and surgery, however the 5-year survival rate remains poor at 60-70%.

Chemoresistance and recurrence of osteosarcoma are the two major challenges confronting physicians today. Tumour necrosis following chemotherapy is an important prognostic indicator and is evaluated at the time of resection. Unfortunately however, 30% of patients prove resistant to current chemotherapeutic agents. Additionally, despite clear margins being demonstrated at the time of resection, approximately half of all patients will present at a later date with disease recurrence, commonly in the form of pulmonary metastases. The inability of chemotherapy to eradicate micrometastatic disease may be responsible for disease progression in these patients.

In order to improve patient prognosis we need to earlier identify and treat those subgroups of patients who are likely to present with chemoresistance or disease recurrence. We need to develop novel strategies for these patients that will enhance current chemotherapy regimes.

Pigment Epithelium-Derived Factor (PEDF)
– From Eye to Tumor Biology

Pigment epithelium-derived factor (PEDF) is a physiologic molecule that is showing promise as a targeted therapy for osteosarcoma. Our laboratory is focused on applying PEDF as an adjunct to existing therapies for osteosarcoma, however PEDF was first discovered in 1991 as a factor secreted in the human fetal eye.¹ At this time, PEDF was shown to be capable of promoting differentiation of retinoblastoma cells.² PEDF is now known to be the most potent anti-angiogenic factor, seven times more potent than endostatin, and has been implicated in the pathogenesis of angiogenic eye disease.³

PEDF’s role as an anti-cancer agent followed from its characterization as an anti-angiogenic.⁴ PEDF is able to target a tumours supporting vasculature, however direct effects on tumour cells have also been demonstrated. Inhibition of tumour cell proliferation is multifaceted and involves the induction of apoptosis and the inhibition of cell cycling.⁵ So far, PEDF’s direct and indirect effects have been demonstrated for malignancies including prostate, breast and pancreatic carcinomas, melanoma and glioma.⁴

Applying PEDF to Osteosarcoma

The role of PEDF in osteosarcoma pathogenesis and treatment has been the major focus of our laboratory since 2002. Our experience with PEDF first began with an examination of the role of cartilage-derived anti-angiogenic factors at the growth plate. Using immunohistochemistry and in situ hybridisation we previously showed that PEDF expression is largely resistricted to the avascular resting, proliferative and upper hypertrophic layers of the growth plate. These are precisely the regions that are consistently resistant to osteosarcoma invasion.^6,7

We next demonstrated PEDF’s direct and indirect effects on osteosarcoma. PEDF inhibits proliferation, migration and invasion of osteosarcoma cells in vitro, effects that can be replicated in vivo.⁸ The inhibition of cell proliferation relies on the induction of apoptosis and the inhibition of cell cycling. We tested PEDF in an in vivo orthotopic model of osteosarcoma. SaOS-2 osteosarcoma cells were treated with PEDF and primary osteosarcoma was induced by intraosseous injection in nude mice. PEDF restricted growth of primary tumours as well as the occurrence of metastatic lesions.^9-11

Recent Results

The Liddy Shriver Sarcoma Initiative has generously provided funding to further characterise the therapeutic role of PEDF for osteosarcoma. We have evaluated PEDF and two peptide derivatives using a model of established orthotopic osteosarcoma. This study stands apart from our previous in vivo work in that it is the first time that PEDF administration has been delayed until after the macroscopic appearance of primary tumours. The study was designed to replicate human presentation, progression and treatment of osteosarcoma. We also sought to define a role for PEDF as a dose-reducing agent for concurrent doxorubicin therapy.

Osteosarcoma was induced by the intraossesous injection of SaOS-2 osteosarcoma cells. Primary tumours became macroscopically apparent at day 20 following inoculation and treatment protocols were initiated at this time. An implanted intra-peritoneal osmotic pump delivered PEDF, or one of two PEDF-derived peptides. Doxorubicin was administered by intra-peritoneal injection.

PEDF administration as a single agent resulted in a dramatic reduction in tumour volume by the study end-point. This inhibitory effect was also observed for both tested peptide derivatives. Extensive osteolysis and soft tissue extension was seen on the plain radiographs of control animals.

Tumour necrosis following chemotherapy is an important prognostic indicator for patients with osteosarcoma. We evaluated tumour necrosis on haematoxylin and eosin stained sections of orthotopic tumours. While tumour necrosis ranged from 9% to 57% of the sections examined, treatment groups were found to be statistically equivalent. TUNEL staining was then performed as a marker of apoptosis, with prominent staining particularly in PEDF treated tumour sections.

Our previous in vivo studies have showed PEDF to be capable of restricting the development of pulmonary metastases, however on this occasion we were unable to replicate this finding. There was no significant difference in the number of pulmonary metastases observed between treatment groups. One explanation for this is that by delaying PEDF treatment until after the macroscopic appearance of primary tumours, we have enabled all animals to advance to an equivalent stage of disease by the study endpoint. Pulmonary metastases in untreated animals, however, were found to be further advanced and larger in size. To better characterize PEDFs effects on the development of pulmonary metastases, we should aim to evaluate pulmonary metastases with real-time imaging. This would improve the clinical relevance of our orthotopic model and potentially allow a differential effect on metastasis to be observed.

In addition to confirming the therapeutic effects of PEDF and its peptide derivatives for osteosarcoma, we have been able to show a synergistic effect when PEDF is co-administered with doxorubicin. Doxorubicin is a commonly used chemotherapeutic for osteosarcoma, however its use is often limited by side effects such as cardiac and renal toxicity. We compared a combined doxorubicin-PEDF treatment protocol to a single agent regime with high dose doxorubicin. The combined therapy achieved an equivalent degree of tumour suppression as the single agent higher dose doxorubicin treatment. Furthermore, tissues obtained from animals that received the combined PEDF-doxorubicin treatment showed none of the features of doxorubicin therapy that were evident when the high dose of doxorubicin was given alone. This suggests a possible dose-reducing role for PEDF when given with doxorubicin.

The final objective of our in vivo studies was to perform the first thorough evaluation of therapeutic safety for PEDF as an anti-cancer agent. We collected serum, skin, hearts, lungs, and small intestine specimens for all treated animals and found no evidence of toxicity related to PEDF administration.

The biological effects of PEDF in the treatment of osteosarcoma have been well characterised by our in vivo studies thus far. These studies provide grounds on which to begin closer study of the molecular pathways employed by PEDF. Little is currently known of the molecular mechanisms PEDF employs to achieve its effects against osteosarcoma. Using two human osteosarcoma cell lines we have been able to show that PEDF induces multiple mediators for the induction of apoptosis and inhibition of cell cycling.

PEDF regulates cell proliferation by both inducing apoptosis and restricting cell cycling. We performed TUNEL assays on both SaOS-2 and SJSA-1 osteosarcoma cell lines to evaluate apoptosis, and immunocytochemistry for Ki67 to evaluate cell cycling. A significant increase in apoptotic cells, and reduction in Ki67 staining, was observed with treatment with PEDF.

We next perfomed a preliminary survey for PEDF-induced mediators of apoptosis, cell cycling and cell differentiation. Using cultured SaOS-2 and SJSA-1 osteosarcoma cell lines, results suggest that PEDF interacts with a number of mediators across multiple tumorigenic pathways. Caspases, Bcl-2, Bmf, Bax, PARP-1, c-Flip, c-Fos, Puma, NFATc1, JNK, p73, Chk1, Akt, p-Akt and the uPA system have been identified as being regulated by PEDF. While these results are encouraging, further work is needed to characterise possible feedback loops and the upstream or downstream nature of these relationships.

Significance and Future Directions

The impact of osteosarcoma on both patients and the community is great. Current treatment regimes are often prolonged, intensive and disabling, with chemoresistance and disease recurrence particularly challenging. We have successfully shown here that PEDF is capable of inhibiting the growth of osteosarcoma in a clinically relevant murine model. PEDF may be used both as a sole agent and in combination with doxorubicin. Our data suggests that the true potential of PEDF is as a targeted therapy for osteosarcoma, with the added benefit of reducing morbidity of conventional therapies. Our evaluation of SVO2 and SVO3, two PEDF-derived peptides, when combined with our in vitro studies, substantially advances our understanding of the molecular basis of PEDFs anti-osteosarcoma activity.

While our results with PEDF are encouraging, a number of areas for future investigation have emerged. Disease recurrence in the form of pulmonary metastases is the cause of death for many patients with osteosarcoma. We have not yet tested the effect of PEDF on already developed pulmonary metastases. We are currently looking at ways to improve the clinical relevance of our orthotopic model, incorporating real-time imaging of metastatic lesions, so that we may test PEDF in more advanced stages of disease. Additionally, novel and clinically applicable strategies for PEDF delivery require evaluation and may involve the direct delivery of PEDF to sites of metastasis.

V7N3 ESUN Copyright © 2010 Liddy Shriver Sarcoma Initiative.

By Peter Choong, MD and Crispin Dass, PhD

Introduction to Osteosarcoma and its Treatment

Osteosarcoma (OS) is the most common malignancy of bone and is the second highest cause of cancer-related deaths in the paediatric age group. 465 new cases of OS occurred in Australia between 1995-99, an increase of 15% in comparison to a similar period a decade earlier (Australian Institute of Health and Welfare, Cancer in Australia, 2001). The burden to patients and the community is high with disabling surgery, prolonged rehabilitation and treatment that is expensive, protracted, intensive and requiring inpatient monitoring. Importantly, an average of 17 life-years per patient is lost due to sarcomas, compared to 6.5 for bowel, lung and breast cancer (Canstat Victoria, 2001 data). This contributes to a disproportionate cost to the individual and community and targets this type of cancer as an age-related phenomenon which is now a major public health issue.

While surgery remains the mainstay of OS treatment, chemotherapy has delivered significant improvements in survival. Despite this, fatality from recurrent disease still approaches 1/3 of those afflicted. Intensification of chemotherapy has failed to demonstrate significant survival improvement in patients with aggressive relapse highlighting the need for novel therapies that increase the effectiveness while reducing the toxicity of treatment.¹ The importance of translating basic science to clinical practice is becoming a priority for government (Victorian Cancer Agency, 2006), and funding bodies (NHMRC). Identifying targeted therapies that aim to enhance the effect or reduce the toxicity of conventional chemotherapy when treating primary or metastatic OS is a primary goal of this project.

Cartilage is a Barrier to Osteosarcoma Growth

OS arises most commonly around the growing ends of long bones (metaphysis) adjacent to the growth plate cartilage (GPC) suggesting a disruption of physiologic control in the region earmarked for normal proliferation. Despite the destructive capacity of OS, cartilage acts as a strong barrier against invasion and is only penetrated as a late local event. We were the first to link the impediment to advance of OS through the GPC with regulator(s) of bone maturation and the state of cartilage lacking an adequate supply of blood vessels.² In analysing the molecular components of the GPC, we identified for the first time the presence of a protein called pigment epithelium-derived factor (PEDF), the strongest physiologic inhibitor of angiogenesis known, in the part of the GPC lacking blood vessels.³

PEDF has been shown to be more than twice as potent as angiostatin, and more than seven times as potent as endostatin, two endogenous molecules that are potently anti-angiogenic in their own right.⁴ We demonstrated that OS regularly stopped advancing when it met the layer of the GPC at which PEDF protein was abundant (Figure 1). In contrast, OS grew in that part of bone where a protein called vascular endothelial growth factor (VEGF), a potent pro-angiogenic factor, was present at high levels.³ The interface between PEDF and VEGF was the point of cessation of tumor advancement suggesting a balance between opposing pro-and anti-angiogenic forces and this is consistent with the fundamental role of angiogenesis in tumor biology that is also applicable to OS.

Our Data on PEDF and Osteosarcoma

We have compiled a considerable amount of published data that demonstrates the effectiveness of PEDF in inhibiting migration, invasion and proliferation of OS cells in vitro (Figure 2) and these actions can be replicated in vivo.

Recently, we reported that by forcibly expressing (overexpressing) PEDF in our OS model, tumor growth and metastasis were profoundly inhibited (Figure 3). A similar effect was also noted when we exposed developing tumor cells to recombinant PEDF protein (rPEDF). To improve the effectiveness of our strategy, we isolated shorter derivatives of the PEDF protein which contained the regions responsible for anti-angiogenic and anti-proliferative/apoptotic (programmed cell death) activity. By exposing cells in vitro and tumor in vivo, significant antitumoral activity was seen with the less expensive peptides.

To see if apoptosis is associated with PEDF treatment, we treated SaOS-2 cells with rPEDF and noted a dose-related increase in apoptosis (Figure 4). A specific inhibitor of PEDF, an antibody, inhibited the apoptotic effect of rPEDF noted in earlier experiments, confirming that the pro-apoptotic effect was indeed due to PEDF activity. Interrogating the Fas/Fas ligand cell death pathway would be a focus of this project as apoptosis by the induction of Fas ligand by PEDF has been alluded to before.⁷ In addition, we will also look at other cell death pathways such as tumor necrosis factor (TNF)/TNF receptor and TNF-related apoptosis-inducing ligand (TRAIL) and receptors.

There are a number of important interactions between PEDF, VEGF and the latter’s receptor. Reports infer a regulatory action of PEDF on VEGF as well as an inverse relationship between the two.⁸ More recent reports suggest a role for PEDF in regulation of VEGF gene expression.¹¹ We have demonstrated the ability of rPEDF to inhibit angiogenesis by disrupting the expression of VEGF in OS cells (Figure 5). We would like to further tease out the mechanistic interplay between PEDF and VEGF.

To date, experimental administration of PEDF to tumors in vivo has been either via direct injection of rPEDF into the tumor or gene transfer via plasmid DNA or viruses in animal models of disease.⁵ Viral vectors suffer from the risk of de novo cancer initiation via recombination within the patient’s DNA. Plasmid vectors are safer, but are often hindered by low efficiency of gene expression. The use of recombinant protein is attractive in that it avoids the above difficulties and may be produced in bulk via established industrial methods, is easier to handle, and thus, facilitates compliance by treating teams and patients. Recombinant proteins, however, are often unstable and inherently susceptible to biological degradation, while high molecular weight proteins are often associated with low tissue penetration, poor availability at the target site, and uptake into target cells. Production of larger proteins is frequently hampered by high costs and difficulties with handling and storage as has been the case with endostatin, a 20 kDa polypeptide. As PEDF is a 50 kDA protein, it is potentially more vulnerable to the constraints facing endostatin.

A more rational and advantageous approach would be to use active smaller peptide derivatives of PEDF that maintain bioactivity. We generated four 25-mer synthetic peptides (SVO 1-4) derived from parent PEDF corresponding to functional regions, which suppressed OS cell proliferation, inhibited cell invasion, increased cell adhesion to type-1 collagen and reduced VEGF protein.¹² SVO2 and SVO3 demonstrated consistently the highest activity in vitro, and in vivo exhibited a greater than 30% reduction in tumor volume by day 32 and a marked reduction in pulmonary metastases (Figure 6). These results demonstrate successful translation of our initial basic work to a treatment strategy for OS. The implementation of systemic delivery of peptide derivatives will be a major initiative of this project.

Significance of the Project

Our research project thus combines basic science investigations with a prominent translational element. We aim to build on the preceding data by testing rPEDF and its cheaper derivatives on established primary and metastatic OS in our model, individually and in combination with doxorubicin. This series of experiments will analyse the important role of rPEDF and its ability to reduce the toxicity of doxorubicin when used in combination. The molecular postulates that we believe to underpin the effectiveness of PEDF and which we shall explore include the regulation of pro-angiogenic factors and the induction of apoptosis by PEDF. The main focus in our lab is to progress our PEDF research findings towards clinical evaluation and hopefully provide the first instance of successful molecular therapy for OS using endogenous biologicals.

V5N4 ESUN Copyright © 2008 Liddy Shriver Sarcoma Initiative.

Grant Funding

The Liddy Shriver Sarcoma Initiative funded this $50,000 study in August 2008. The grant was made possible, in part, by a generous gift from the Una O'Hagan family in loving memory of her son, Sean Keane.