The Role of Cytoplasmic p27 in Metastatic Osteosarcoma

By Yiting Li, MD, PhD and Tsz-Kwong Man, PhD

Abstract

Osteosarcoma is the most common malignant bone tumor in children. Recent studies have shown that the tumor suppressor p27 protein promotes metastasis when it is mislocalized from the nucleus to the cytoplasm in some solid tumors; however, the localization of the p27 protein in osteosarcoma and its role in metastasis is largely unknown. In this study, we found that many human osteosarcoma cases harbored cytoplasmic staining of p27. Immunohistochemical staining showed that the p27 protein was localized in the nuclei of non-metastatic osteosarcoma cell lines, but mislocalized in the cytoplasm of their metastatic sublines. In addition, overexpression of cytoplasmic p27 in a non-metastatic osteosarcoma cell line significantly increased tumor cell migration and invasion in vitro. Mechanistic investigations further revealed that phosphorylations of two amino acids on the p27 protein were essential for cell migration and invasion. We found that rapamycin, an mTOR inhibitor, reversed the protein trafficking of p27 from the cytoplasm to the nuclei of metastatic osteosarcoma cells, suggesting that p27 protein trafficking is dependent on mTOR signaling. In summary, our results indicate that a novel relationship exists between p27 mislocalization and the metastatic potential of osteosarcoma cells. Targeting p27 may have therapeutic potential in the treatment of metastasis of osteosarcoma.

An Introduction to Osteosarcoma and p27

Osteosarcoma is the most common malignant bone tumor in children and young adults. Despite the advent of chemotherapy, metastasis continues to be the major cause of death in osteosarcoma patients. The survival rate of osteosarcoma patients with metastasis continues to be poor.¹ In order to develop targeted therapies that can improve outcomes for these patients, it is important to identify the key genes or proteins that promote metastasis in osteosarcoma.

p27 is a nuclear protein that inhibits cell cycle progression and acts as a tumor suppressor. Unlike the other tumor suppressor genes, p27 is rarely mutated in human cancers.² Recent studies have found that the p27 protein is able to promote metastasis in some solid tumors when it is mislocalized in the cytoplasm.^3-5 The goals of this study were to elucidate the subcellular localization of the p27 protein in osteosarcoma and its role in metastasis. To achieve these goals, we first examined the expression and subcellular localization of the p27 protein in an osteosarcoma tissue microarray as well as three pairs of isogenic metastatic osteosarcoma cell lines. Then, p27 was overexpressed in the cytoplasm of a commonly used non-metastatic osteosarcoma cell line and the effects on cell motility and invasion were examined. We further tested to see if aberrations in various phosphorylation sites on the p27 protein affect cell motility. The relationship between p27 and mTOR, which is known to promote metastasis of osteosarcoma, was also studied.

p27 protein is encoded by CDKN1B. In the nucleus, it inhibits the formation of cyclin E-Cdk2 complex and arrests cell cycle transition from G0 phase to S phase. Therefore, p27 has been regarded as a tumor suppressor gene for decades. Interestingly, recent studies found an independent metastasis-promoting function of p27. In cancer cells, the p27 protein can be phosphorylated by AKT or the other kinases, and the phosphorylation prohibits its entry to the nucleus. The cytoplasmic mislocalized p27 interacts with cytoskeleton proteins and regulates cell motility and metastasis as indicated in a variety of tumors.3-5

Recent Results

To determine the subcellular localization of the p27 protein in human osteosarcoma cases, we performed p27 immunohistochemical staining on an osteosarcoma tissue microarray provided by the Children’s Oncology Group. There are 63 total osteosarcoma cases in the microarray. Strikingly, a majority of the OS cases (82%) display cytoplasmic staining of p27, suggesting that p27 is frequently mislocalized in the cytoplasm of human osteosarcoma cases (Figure 1A). To test if the protein mislocalization is associated with metastatic osteosarcoma, we examined three pairs (metastatic and non-metastatic) of isogenic osteosarcoma cell lines. We found that p27 was located in nuclei of the parental non-metastatic cell lines, but the protein was retained in the cytoplasm in all three metastatic sub-lines (Figure 1B). This observation suggests that the cytoplasmic mislocalization of p27 may play a role in the metastasis of osteosarcoma.

We then tested to see if cytoplasmic p27 was functionally involved in the metastatic process of osteosarcoma. We used a recombinant construct that fused a Nuclear-Exporting Sequence (NES) tag with the p27 gene, so that the p27 protein, which normally resides in the nucleus, will be exported to the cytoplasm. In addition, the cyclin-CDK binding domain of the p27 gene that controls cell cycle progression was mutated in the construct (ck-) to ensure that the effect of this NES-p27 mutant is independent from its known function on cell cycle regulation. Then, the NES-p27 ck- construct was stably expressed in the SaOS-2 cells, which is a commonly used non-metastatic/lowly metastatic osteosarcoma cell line (Figure 2A). Using a transwell assay, we demonstrated that the expression of cytoplasmic p27 significantly increased the osteosarcoma cells’ ability of migration and invasion (Figure 2B). Since cell migration and invasion are important characteristics of metastasis, p27 may promote metastasis in osteosarcoma when it is mislocalized in the cytoplasm.

Nuclear-Exporting Sequence (NES) refers to a short sequence containing 4 hydrophobic amino acids that guides the export of a nuclear protein to cytoplasm through nucleus pores.

The CDK-binding domain of the p27 protein binds to the ATP-binding site of Cyclin-Dependent Kinase (CDK), and prohibits its activation by Cyclin. The CDK-binding domain is required for the cell cycle regulation effect of p27.

The transwell assay is widely used to evaluate the ability of cells to migrate or invade. To perform the assay, cells are seeded on the surface of a permeable membrane inside a well. If appropriate, solution that attracts cell migration and invasion can be placed underneath the membrane and the well. After an incubation period, the cells retained on the surface are removed and those transported (migrated or invaded) through the membrane are stained and counted.

p27 is subject to complex post-translational regulations, such as protein degradation and phosphorylation.^6-10 Some of the phosphorylation sites are known to be important to the subcellular localization of the protein.^8-10 To determine if the phosphorylation sites on the p27 protein are important for promoting cell motility, we generated mutations on three key phosphorylation sites in the NES-p27 ck- mutant (Figure 3A). Then, these mutated constructs were stably expressed in SaOS-2 cells, and the effects of the mutations on cell migration were examined. As shown in Figure 3B, the mutations on Serine-10 and Threonine-198 residues of the p27 protein abolished the promoting effect of NES-p27 ck- on cell migration, while the mutation on Threonine-157 residue has no effect on cell migration. This result suggests that the Serine-10 and Threonine-198 residues are important to the metastatic potential of the osteosarcoma cells even if the p27 protein is mislocalized in the cytoplasm. Targeting these two phosphorylation sites may have therapeutic significance.

Previous evidence suggests that both Protein kinase B (AKT)9,10 and Mammalian target of Rapamycin (mTOR)12 can phosphorylate the p27 protein. Since the mTOR pathway has been implicated in the metastasis of osteosarcoma, we, therefore, further dissected the relationship between the p27 mislocalization and the mTOR pathway. We performed immunohistochemistry to examine the subcellular localization of p27 in the presence of an mTOR inhibitor, rapamycin. Using 40nM rapamycin, we found that the number of LM7 cells (a metastatic subline of SaOS-2) with cytoplasmic p27 decreased. No effect was observed in the control SaOS-2 cells. Further increase of rapamycin concentration resulted in extensive cell death, suggesting the reversal of the trafficking of cytoplasmic p27 into the nuclei of the metastatic osteosarcoma cells may restore the cell cycle inhibition effect of p27 and cause cell death. This result suggests that cytoplasmic mislocalized p27 is downstream of mTOR, and inhibition of mTOR signaling has an anti-tumor effect on the metastatic osteosarcoma cells.

AKT is a serine/threonine kinase that regulates cell growth, survival, and migration. It has been implicated as a major factor in a variety of cancer types, including osteosarcoma.

mTOR (mammalian target of rapamycin) is a protein kinase that regulates cell growth, survival, and motility. Recent studies found that the mTOR pathway is involved in the metastasis of osteosarcoma. Treatment with rapamycin, the inhibitor of mTOR, significantly reduces lung metastasis in animal models. ¹¹

Summary

From the results obtained from this study, we found that p27 mislocalization is a frequent event in osteosarcoma cases. Cytoplasmic mislocalized p27 can increase cell migration and invasion. We also identified two phosphorylation sites (Serine-10 and Threonine-198) on the p27 protein that are important for the tumor cell migration. Inactivation of the residues may have a therapeutic value in metastatic osteosarcoma. Lastly, we demonstrated that p27 trafficking is affected by an mTOR inhibitor. This suggests that p27 is downstream of the mTOR signaling, and mTOR inhibitors may have a therapeutic effect for metastatic osteosarcoma by affecting the p27 mislocalization. In summary, we have demonstrated the metastasis promoting effect of cytoplasmic p27 in osteosarcoma. Further investigations of the results from this study may lead to development of a novel targeted therapeutic strategy for metastatic osteosarcoma.

Future Direction

We are in the process of confirming the in vivo metastasis-promoting effect of cytoplasmic p27 using a mouse model. We are also investigating the relationship between p27 mislocalization and the RHOA pathway, and evaluating compounds/genes that can inhibit the p27-induced metastasis. The long-term goal is to elucidate the effect of p27 mislocalization in osteosarcoma, so that a novel therapeutic strategy can be developed to improve the outcome of osteosarcoma patients with metastasis.

By Yiting Li., MD, PhD^1,2
and Tsz-Kwong Man, PhD^1,2,3
¹Texas Children’s Cancer and Hematology Centers, Texas Children’s Hospital, Houston, TX
²Department of Pediatrics
³Dan L. Duncan Cancer Center
Baylor College of Medicine, Houston, TX

References

1. Link MP, Eilber F. Osteosarcoma. In: Eilber F, Poplack D, editors. Principles and Practice of Pediatric Oncology. 3rd ed. Philadelphia: Lippincott-Raven Publishers; 1997. p. 889-920.

2. Ponce-Castaneda MV, Lee MH, Latres E, Polyak K, Lacombe L, Montgomery K, Mathew S, Krauter K, Sheinfeld J, Massague J. p27Kip1: Chromosomal mapping to 12p12-12p13.1 and absence of mutations in human tumours. Cancer Res 1995; 55:1211-4.

3. Singh SP, Lipman J, Goldman H, Ellis FH Jr, Aizenman L, Cangi MG, Signoretti S, Chiaur DS, Pagano M, Loda M. Loss or altered subcellular localization of p27 in Barrett's associated adenocarcinoma. Cancer Res 1998; 58:1730-5.

4. Ciaparrone M, Yamamoto H, Yao Y, Sgambato A, Cattoretti G, Tomita N, Monden T, Rotterdam H, Weinstein IB. Localization and expression of p27KIP1 in multistage colorectal carcinogenesis. Cancer Res 1998; 58:114-22.

5. Denicourt C, Saenz CC, Datnow B, Cui XS, Dowdy SF. Relocalized p27(Kip1) tumor suppressor functions as a cytoplasmic metastatic oncogene in melanoma. Cancer Res 2007; 67:9238-43.

6. Loda M, Cukor B, Tam SW, Lavin P, Fiorentino M, Draetta GF, Jessup JM, Pagano M. Increased proteasome-dependent degradation of the cyclin-dependent kinase inhibitor p27 in aggressive colorectal carcinomas. Nat Med 1997; 3:231-4.

7. Michelle D. Larrea, Seth A. p27 as Jekyll and Hyde regulation of cell cycle and cell motility. Cell Cycle 2009; 8: 3455-3461.

8. Besson A, Gurian-West M, Chen X, Kelly-Spratt KS, Kemp CJ, Roberts JM. A pathway in quiescent cells that controls p27Kip1 stability, subcellular localization, and tumor suppression. Genes Dev. 2006 20:47-64.

9. Motti ML, De Marco C, Califano D, Fusco A, Viglietto G. Akt-Dependent T198 phosphorylation of Cyclin-Dependent Kinase Inhibitor p27(kip1) in breast Cancer. Cell Cycle. 2004; 3:e89-e95.

10. Short JD, Houston KD, Dere R, Cai SL, Kim J, Johnson CL, Broaddus RR, Shen J, Miyamoto S, Tamanoi F, Kwiatkowski D, Mills GB, Walker CL. AMP-activated protein kinase signaling results in cytoplasmic sequestration of p27. Cancer Res. 2008; 68:6496-506.

11. Wan, X., Mendoza, A., Khanna, C., Helman, L. J. Rapamycin inhibits ezrin-mediated metastatic behavior in a murine model of osteosarcoma. Cancer Res. 2005; 65, 2406-11.

12. Hong F, Larrea MD, Doughty C, Kwiatkowski DJ, Squillace R, Slingerland JM. mTOR-raptor binds and activates SGK1 to regulate p27 phosphorylation. Mol Cell. 2008; 30:701-11.

Targeting the tumor-associated antigen p27kip1 in metastatic osteosarcoma

By Yiting Li, MD, PhD and Tsz-Kwong Man, PhD

Abstract

Osteosarcoma is the most common bone cancer in children and adolescents, which accounts for approximately 60% of bone cancers in the first two decades of life. With the advent of chemotherapy, the survival rate of osteosarcoma patients has reached 60-70%. However, if the tumor cells metastasize to other parts of the body, a majority of the affected patients will succumb to the disease. The major challenge of the field now is to develop a novel and targeted approach that can facilitate the treatment of patients with disease spread. Previous studies have demonstrated that the body’s immunity generates specific autoantibodies, which target proteins that are important to the progression and malignancy of tumor cells. Using this autoantibody approach, we recently identified the autoantibody for the cell cycle inhibitor p27kip1, which was significantly higher in patients with metastatic osteosarcoma. Functional analysis further demonstrated the protein was mislocalized in metastatic osteosarcoma cells. p27kip1 (encoded by CDKN1B) is an atypical tumor suppressor that regulates G0 to S phase transitions by inhibiting the cyclin E-Cdk2 complex. Recent studies also suggest that the cytoplasmic mislocalization of p27kip1 can promote metastasis in cancers. In addition, the mTOR pathway has been implicated in metastatic osteosarcoma. Elucidating the role of p27kip1 in metastatic osteosarcoma and, particularly, its relationship with the mTOR pathway, will enhance our understanding of the metastasis mechanism in osteosarcoma and the potential use of this cell cycle inhibitor as a therapeutical target.

Current Treatment of Osteosarcoma

Osteosarcoma is the most common malignant bone tumor in children and young adults. The standard of care for osteosarcoma involves four steps:

Diagnosis by an initial biopsy
Two to three courses of multi-agent chemotherapy
Definitive surgery to resect the tumor
Post-operative chemotherapy based on the patient’s response to pre-operative chemotherapy

Patients who have a good response to pre-operative chemotherapy will receive the same chemotherapy postoperatively. Poor responders will receive intensified chemotherapy in an attempt to counter the relative chemoresistance of the tumor. However, the promise of using the response to pre-operative chemotherapy to customize post-operative therapy has not been kept since it has failed to improve the osteosarcoma disease-free survival in high-risk patients who are poor responders. Despite several multi-center clinical trials in both Europe and North America that have taken advantage of the pre-operative chemotherapy strategy, the overall survival rate has not improved significantly during the past 30 years.¹ One way to improve the overall outcome is to customize therapy based on prognostic factors. This could improve survival in those patients predicted to have a poor prognosis by intensifying the use of known effective agents. It might also allow the reduction of dose-related toxicity while maintaining good survival rates for patients predicted to have a good prognosis. The above strategy is predicated on our ability to identify groups of patients with either a good or a poor prognosis and stratify them for a risk-based, personalized treatment.

Identification of Biomarkers in Osteosarcoma

In the effort to identify prognostic factors in osteosarcoma, we have used microarray and bioinformatic approaches to identify an expression signature, which could predict osteosarcoma chemoresistance at the time of diagnosis (Figure 1A).² In this study, we developed a molecular classifier that can predict the likelihood of developing chemoresistance at the time of initial diagnosis, i.e. before the pre-operative chemotherapy. We further hypothesize that the tumor tissues isolated from the definitive surgery contain enriched resistant tumor cells that can be used to identify the chemoresistance signature. Therefore, we used the expression profiles from tumor specimens collected at definitive surgery to train our classification algorithms and found that the classifier developed by the SVM algorithm performed the best in a leave-one-out cross validation. Thus, we further tested if this SVM classifier can be used to predict the chemotherapy response using the specimens collected at initial diagnosis. The results showed that the SVM classifier exhibits a high level of accuracy (83%) in predicting the poor responders using initial biopsy specimens. In addition to expression profiling, we also described the genomic aberrations in osteosarcoma and identified the associated genes in the highly amplified and deleted regions (Figure 1B).⁴

Metastasis is the Major Adverse Prognostic Factor in Osteosarcoma

In addition to the response to chemotherapy, another major prognostic factor in osteosarcoma is the development of metastasis. In fact, clinically detectable metastasis at initial diagnosis is the most consistent adverse prognostic factor in osteosarcoma.⁵ Approximately 20% of osteosarcoma patients have clinically detectable metastasis at the time of diagnosis.¹ The lung is the most common site for osteosarcoma metastasis.⁶ Lung metastasis has a major impact on prognosis for osteosarcoma patients. Even with the multi-agent chemotherapy used in current treatment protocols, the progression-free survival rate for patients with metastatic osteosarcoma is still low (< 20%) (Figure 2).^7-8 This suggests that the current treatment regimen is not effective for patients with clinically detectable metastasis. Despite the small percentage of patients with overt metastases at diagnosis, over 80% of patients with osteosarcoma treated with excision alone are eventually diagnosed with metastases. This observation suggests that many patients with this disease harbor micrometastases at diagnosis or before the definitive surgery.⁹ A recent retrospective study of 280 patients with metastatic osteosarcoma who had metastases detected at various times during treatment showed that the time of metastasis identification is an important prognostic factor.¹⁰ The 5-year survival rates of patients who had lung metastasis detected during pre-operative and post-operative chemotherapy were 0% and 6%, respectively compared with 18% for those who were detected at initial diagnosis. Therefore, it is important to develop a targeted and more effective therapy to treat patients with metastasis when the chance of cure is still high.

The Use of Tumor-Associated Antigens to Identify Important Targets in Osteosarcoma

Since proteins ultimately determine the phenotype of the cell and genomic profiling is not particularly suitable for identifying circulating biomarkers in peripheral blood, we have recently extended our osteosarcoma studies to proteomic analysis.^11-12 Using proteomic profiling, we have demonstrated that plasma contains protein peaks that can be used to distinguish malignant osteosarcoma from benign osteochondroma.¹² Due to low quantities of the tumor-associated proteins in peripheral blood and the lack of a protein amplification method, current proteomic approaches are often not sensitive for detecting the tumor-specific proteins in plasma. Although there is no in vitro protein counterpart to the polymerase chain reaction for amplification of nucleic acids, the body’s immunogenic response produces large amount of antibodies that are specifically recognized tumor-associated antigens.¹³ Cancer patients often produce autoantibodies to tumor-associated antigens that are overexpressed, post-translationally modified, aberrantly cleaved, or aberrantly localized in tumor cells (Figure 3).

Given that tumor-associated autoantibodies are considered reporters of the immune response to a developing tumor, they are excellent candidates for cancer biomarkers for early detection. Autoantibodies against tumor-associated antigens have been shown to be useful in both cancer diagnosis and prognosis.¹⁴ In addition, characterization of these tumor-associated antigens will likely identify cellular targets that are important for the tumor. Brichory et al reported autoantibodies against annexin I and annexin II and later against PGP 9.5 in 30%-50% of patients with lung adenocarcinomas with high specificity.^15,16 Imafuku et al have summarized autoantibodies reported for various cancers, including calreticulin, MUC1, p53, and Rad51.¹⁷

A New High-Throughput Platform for the Identification of Tumor-Associated Antigens

Based on this tumor-associated antigen phenomenon, we have recently used a new protein microarray platform ― ProtoArray Human Protein Array (ProtoArray HPA), to identify autoantibodies and tumor-associated antigens in patients with metastatic osteosarcoma. Each ProtoArray HPA contains 8,268 purified human proteins that cover many known cancer-related protein and functional classes, including protein kinases, transcription factors, membrane proteins, nuclear proteins, signal transduction, secreted proteins, cell communication, metabolism, cell death and proteases/peptidases (Figure 4). This new approach has several advantages over traditionally used autoantibody-screening approaches.

First, unlike the gel-based approach, ProtoArray HPA can detect basic proteins and membrane proteins. Second, this array system consists of thousands of specific and purified proteins that prevent the problem associated with the peptide phage surface display approach of detecting non-biological mimotopes.¹⁴ In addition, many tumor-associated antigen arrays reported in the literature contain only several hundreds of specific antigens,¹⁸ ProtoArray HPA has a much larger protein coverage than those arrays. Even with the 2-dimensional gel system that can typically visualize 1000-2000 spots on a gel, the ProtoArray HPA can screen from fourfold to eightfold more proteins and their identities are already known. Thus, no downstream protein identification is necessary. In contrast, phage display is based on multiple enrichment steps; thus, this assay is only qualitative, exhibits bias toward abundant antibodies, and may miss important autoantibodies during the enrichment. A recent study of using the Protoarray HPA in ovarian cancer has found many new tumor-associated antigens, and many of them were overexpressed in tumor tissues.¹⁹

p27kip1 and Osteosarcoma Metastasis

Using this new microarray platform, we have identified a tumor-associated autoantibody in metastatic osteosarcoma, which is targeting p27kip1. p27kip1 (encoded by CDKN1B) is an atypical tumor suppressor that regulates G0 to S phase transitions by inhibiting the cyclin E-Cdk2 complex (Figure 5).^20,21 Interestingly, it is rarely mutated in human tumors.²² Recent studies have shown that p27kip1 can inhibit or promote cell motility.²³ By binding to a microtubule destabilizing protein, stathmin, p27kip1 has been shown to inhibit the migration of fibrosarcoma cells and normal fibroblasts. Binding of p27kip1 to stathmin can reduce the stathmin binding to tubulin, which causes an increase of microtubule polymerization (Figure 5).²⁴ In contrast, other studies have also shown that the cytoplasmic p27kip1 protein may affect actins and RhoA, which can stimulate cell migration.^25-27 Overexpression of a p27kip1 mutant localizing to the cytoplasm increased melanoma cell motility and metastasis in vivo.²⁸ When a p27kip1 construct with a mutation in Cdks was introduced into p27kip1 null mice, the protein was localized to both nucleus and cytoplasm of the cell. The animals with the mutant construct showed an expansion of tissue progenitor cells and developed tumors in multiple organs.²⁹ Thus, the oncogenic effect of cytoplasmic p27kip1 is independent of Cdks and its effect on cell migration may depend on stathmin or RhoA. This metastatic function of p27kip1 may explain the low frequency of p27kip1 loss in human cancers. However, the role of p27kip1 in osteosarcoma or other pediatric cancers is largely unknown. It is not clear how the signaling of p27kip1 affects its subcellular localization in tumor cells. Some evidence seems to suggest that phosphorylation of p27kip1 by AKT promotes the cytoplasmic mislocalization of p27kip1.³⁰ Interestingly, previous studies have implicated the Ezrin-mTOR pathway in metastatic osteosarcoma³¹; however, the relationship of p27kip1 and the mTOR pathway has not been examined. Since both p27kip1 and mTOR are downstream of PI3K-AKT pathway, it is very likely that the metastatic effect of cytoplasmic p27kip1 may be linked to the mTOR pathway in osteosarcoma (Figure 5). In this study, we propose experiments to determine if cytoplasmic mislocalization of p27kip1 will lead to a promotion of cell migration and invasiveness in osteosarcoma cells. We will also dissect the phosphorylation events on p27kip1 that are critical of its mislocalization. Lastly, we propose to investigate the relationship between p27kip1 and the AKT-mTOR pathway.

Future Direction

Despite the extensive efforts in the field, metastasis is still the leading cause of death in pediatric osteosarcoma patients. Even with multi-agent chemotherapy, a significant portion of patients develop metastasis after initial diagnosis and treatment, suggesting a more effective therapeutic approach is needed to improve the outcome of these patients. Our preliminary studies and studies elsewhere have already suggested that cytoplasmic p27kip1 plays an important role in metastasis in osteosarcoma and other cancers. Understanding the role of this important kinase inhibitor in metastatic osteosarcoma may result in the development of a novel therapeutic approach for this deadly cancer. Future studies will include validating the association of p27kip1 with metastatic osteosarcoma using specimens collected from a multi-institutional osteosarcoma study, the joint European/American Osteosarcoma Study (EURAMOS). We will also evaluate the use of p27kip1 as therapeutic targets in a panel of metastatic osteosarcoma cell lines, and studying the in vivo metastatic effect of cytoplasmic p27kip1 in animal models. Since cytoplasmic p27kip1 has been implicated in other metastatic cancers, the results obtained in this study may be applicable to other pediatric metastatic sarcomas or cancers and may lead to new research directions

Glossary of Terms Found in this Article

A classifier is a supervised machine learning or statistical method for assigning samples into different classes.

Cross validation is an analytical technique of measuring if the results of a classification can be generalized to other independent data sets.

EURAMOS is a randomized trial of osteosarcoma patients from Europe and North America to optimize treatment strategies for resectable osteosarcoma based on histological response to pre-operative chemotherapy.

Microarray is a multiplex technology used in molecular biology and in medicine. It consists of an arrayed series of thousands of microscopic spots of DNA molecules on a glass slide or a similar platform.

Metastasis is the spread of cancer cells from one organ or site to another organ or site.

Peptide phage surface display is a molecular biology method for studying peptide-peptide or protein-protein interactions. The method uses a small virus (phage) that infects bacteria to encode and display the peptide of interest in order to select the interacting peptides.

Polymerase chain reaction (PCR) is a molecular biology technique to amplify DNA molecules through a series of enzymatic reactions.

Support Vector Machine (SVM) is an algorithm used in a classification problem.

By Yiting Li, MD, PhD
and Tsz-Kwong Man, PhD
Texas Children’s Cancer Center
Dan L. Duncan Cancer Center
Baylor College of Medicine

References

1. Link MP, Eilber F. Osteosarcoma. In: Eilber F, Poplack D, editors. Principles and Practice of Pediatric Oncology. 3rd ed. Philadelphia: Lippincott-Raven Publishers; 1997. p. 889-920.

2. Man, T. K., Chintagumpala, M., Visvanathan, J., Shen, J., Perlaky, L., Hicks, J., Johnson, M., Davino, N., Murray, J., Helman, L., Meyer, W., Triche, T., Wong, K. K., Lau, C. C. (2005) Expression profiles of osteosarcoma that can predict response to chemotherapy. Cancer Res 65, 8142-50.

3. Petrilli, A. S., de, C. B., Filho, V. O., Bruniera, P., Brunetto, A. L., Jesus-Garcia, R., Camargo, O. P., Pena, W., Pericles, P., Davi, A., Prospero, J. D., Alves, M. T., Oliveira, C. R., Macedo, C. R., Mendes, W. L., Almeida, M. T., Borsato, M. L., dos Santos, T. M., Ortega, J., Consentino, E. (2006) Results of the Brazilian Osteosarcoma Treatment Group Studies III and IV: prognostic factors and impact on survival. J Clin Oncol 24, 1161-8.

4. Man, T. K., Lu, X. Y., Jaeweon, K., Perlaky, L., Harris, C. P., Shah, S., Ladanyi, M., Gorlick, R., Lau, C. C., Rao, P. H. (2004) Genome-wide array comparative genomic hybridization analysis reveals distinct amplifications in osteosarcoma. BMC Cancer 4, 45.

5. Bielack, S. S., Kempf-Bielack, B., Delling, G., Exner, G. U., Flege, S., Helmke, K., Kotz, R., Salzer-Kuntschik, M., Werner, M., Winkelmann, W., Zoubek, A., Jurgens, H., Winkler, K. (2002) Prognostic factors in high-grade osteosarcoma of the extremities or trunk: an analysis of 1,702 patients treated on neoadjuvant cooperative osteosarcoma study group protocols. J Clin Oncol 20, 776-90.

6. Kaste, S. C., Pratt, C. B., Cain, A. M., Jones-Wallace, D. J., Rao, B. N. (1999) Metastases detected at the time of diagnosis of primary pediatric extremity osteosarcoma at diagnosis: imaging features. Cancer 86, 1602-8.

7. Kager, L., Zoubek, A., Potschger, U., Kastner, U., Flege, S., Kempf-Bielack, B., Branscheid, D., Kotz, R., Salzer-Kuntschik, M., Winkelmann, W., Jundt, G., Kabisch, H., Reichardt, P., Jurgens, H., Gadner, H., Bielack, S. S. (2003) Primary metastatic osteosarcoma: presentation and outcome of patients treated on neoadjuvant Cooperative Osteosarcoma Study Group protocols. J Clin Oncol 21, 2011-8.

8. Bacci, G., Briccoli, A., Rocca, M., Ferrari, S., Donati, D., Longhi, A., Bertoni, F., Bacchini, P., Giacomini, S., Forni, C., Manfrini, M., Galletti, S. (2003) Neoadjuvant chemotherapy for osteosarcoma of the extremities with metastases at presentation: recent experience at the Rizzoli Institute in 57 patients treated with cisplatin, doxorubicin, and a high dose of methotrexate and ifosfamide. Ann Oncol 14, 1126-34.

9. Bacci, G., Lari, S. (2001) Adjuvant and neoadjuvant chemotherapy in osteosarcoma. Chir Organi Mov 86, 253-68.

10. Tsuchiya, H., Kanazawa, Y., Abdel-Wanis, M. E., Asada, N., Abe, S., Isu, K., Sugita, T., Tomita, K. (2002) Effect of timing of pulmonary metastases identification on prognosis of patients with osteosarcoma: the Japanese Musculoskeletal Oncology Group study. J Clin Oncol 20, 3470-7.

11. Man, T. K., Li, Y., Dang, T. A., Shen, J., Perlaky, L., Lau, C. C. (2006) Optimising the Use of TRIzol-extracted Proteins in Surface Enhanced Laser Desorption/ Ionization (SELDI) Analysis. Proteome Sci 4, 3.

12. Li, Y., Dang, T. A., Shen, J., Perlaky, L., Hicks, J., Murray, J., Meyer, W., Chintagumpala, M., Lau, C. C., Man, T. K. (2006) Identification of a plasma proteomic signature to distinguish pediatric osteosarcoma from benign osteochondroma. Proteomics 6, 3426-35.

13. Hanash, S. (2003) Harnessing immunity for cancer marker discovery. Nat Biotechnol 21, 37-8.

14. Wang, X., Yu, J., Sreekumar, A., Varambally, S., Shen, R., Giacherio, D., Mehra, R., Montie, J. E., Pienta, K. J., Sanda, M. G., Kantoff, P. W., Rubin, M. A., Wei, J. T., Ghosh, D., Chinnaiyan, A. M. (2005) Autoantibody signatures in prostate cancer. N Engl J Med 353, 1224-35.

15. Brichory, F. M., Misek, D. E., Yim, A. M., Krause, M. C., Giordano, T. J., Beer, D. G., Hanash, S. M. (2001) An immune response manifested by the common occurrence of annexins I and II autoantibodies and high circulating levels of IL-6 in lung cancer. Proc Natl Acad Sci U S A 98, 9824-9.

16. Brichory, F., Beer, D., Le Naour, F., Giordano, T., Hanash, S. (2001) Proteomics-based identification of protein gene product 9.5 as a tumor antigen that induces a humoral immune response in lung cancer. Cancer Res 61, 7908-12.

17. Imafuku, Y., Omenn, G. S., Hanash, S. (2004) Proteomics approaches to identify tumor antigen directed autoantibodies as cancer biomarkers. Dis Markers 20, 149-53.

18. Casiano, C. A., Mediavilla-Varela, M., Tan, E. M. (2006) Tumor-associated antigen arrays for the serological diagnosis of cancer. Mol Cell Proteomics 5, 1745-59.

19. Hudson, M. E., Pozdnyakova, I., Haines, K., Mor, G., Snyder, M. (2007) Identification of differentially expressed proteins in ovarian cancer using high-density protein microarrays. Proc Natl Acad Sci U S A 104, 17494-9.

20. Morgan, D. O. (1995) Principles of CDK regulation. Nature 374, 131-4.

21. Ekholm, S. V., Reed, S. I. (2000) Regulation of G(1) cyclin-dependent kinases in the mammalian cell cycle. Curr Opin Cell Biol 12, 676-84.

22. Chu, I. M., Hengst, L., Slingerland, J. M. (2008) The Cdk inhibitor p27 in human cancer: prognostic potential and relevance to anticancer therapy. Nat Rev Cancer 8, 253-67.

23. Besson, A., Gurian-West, M., Schmidt, A., Hall, A., Roberts, J. M. (2004) p27Kip1 modulates cell migration through the regulation of RhoA activation. Genes Dev 18, 862-76.

24. Baldassarre, G., Belletti, B., Nicoloso, M. S., Schiappacassi, M., Vecchione, A., Spessotto, P., Morrione, A., Canzonieri, V., Colombatti, A. (2005) p27(Kip1)-stathmin interaction influences sarcoma cell migration and invasion. Cancer Cell 7, 51-63.

25. Nagahara, H., Vocero-Akbani, A. M., Snyder, E. L., Ho, A., Latham, D. G., Lissy, N. A., Becker-Hapak, M., Ezhevsky, S. A., Dowdy, S. F. (1998) Transduction of full-length TAT fusion proteins into mammalian cells: TAT-p27Kip1 induces cell migration. Nat Med 4, 1449-52.

26. McAllister, S. S., Becker-Hapak, M., Pintucci, G., Pagano, M., Dowdy, S. F. (2003) Novel p27(kip1) C-terminal scatter domain mediates Rac-dependent cell migration independent of cell cycle arrest functions. Mol Cell Biol 23, 216-28.

27. Wu, F. Y., Wang, S. E., Sanders, M. E., Shin, I., Rojo, F., Baselga, J., Arteaga, C. L. (2006) Reduction of cytosolic p27(Kip1) inhibits cancer cell motility, survival, and tumorigenicity. Cancer Res 66, 2162-72.

28. Denicourt, C., Saenz, C. C., Datnow, B., Cui, X. S., Dowdy, S. F. (2007) Relocalized p27Kip1 tumor suppressor functions as a cytoplasmic metastatic oncogene in melanoma. Cancer Res 67, 9238-43.

29. Besson, A., Hwang, H. C., Cicero, S., Donovan, S. L., Gurian-West, M., Johnson, D., Clurman, B. E., Dyer, M. A., Roberts, J. M. (2007) Discovery of an oncogenic activity in p27Kip1 that causes stem cell expansion and a multiple tumor phenotype. Genes Dev 21, 1731-46.

30. Morishita, D., Katayama, R., Sekimizu, K., Tsuruo, T., Fujita, N. (2008) Pim kinases promote cell cycle progression by phosphorylating and down-regulating p27Kip1 at the transcriptional and posttranscriptional levels. Cancer Res 68, 5076-85.

31. Wan, X., Mendoza, A., Khanna, C., Helman, L. J. (2005) Rapamycin inhibits ezrin-mediated metastatic behavior in a murine model of osteosarcoma. Cancer Res 65, 2406-11.

Grant Funding

The Liddy Shriver Sarcoma Initiative funded this $50,000 grant in February 2010, and it was covered in a press release by Baylor College of Medicine. The study was made possible, in part, by generous donations made in memory of Sammie Hartsfield (Team Sammie), Brandon Gordon (Brandon's Defense Foundation) and Emma Koertzen, all who lost their lives to osteosarcoma; and by generous donations made in honor of Todd Andrews (Team Sarcoma Bike Tour), Logan Brasic (Soccer 'Round the Clock), and Shannon Ryan (Team Sarcoma Bike Tour).

Plan Figure 1. A. A heat map of a gene expression signature that can predict the response to adjuvant chemotherapy at the time of diagnosis (Ref. 2). Good and poor denotes good and poor responders, respectively. Red and green colors indicate expression levels of the gene. B. The frequency plot of chromosomal aberrations in osteosarcoma (Ref. 4). Green and red colors denote gains and losses, respectively. The highly frequent aberrations are circled.
Plan Figure 2: The survival difference between metastatic (Yes) and non-metastatic osteosarcoma (No) patients (Ref. 3).
Plan Figure 3. The autoantibody reaction in the cancer patients. Necrosis or lysis of tumor cells releases a variety of tumor-associated antigens in the circulation. These antigens are taken up and then presented by the antigen-presenting cells to the T-helper cells. The T-helper cells then induce the differentiation of the B-cells to produce autoantibodies specific to the tumor-associated antigens.
Plan Figure 4: The principle of ProtoArray HPA. The microarray contains thousands of proteins on its surface. The proteins will be recognized and bound by the antibodies in the blood sample. Then a fluorescence-conjugated anti-human IgG is added for the detection of the bound antibodies.
Plan Figure 5. The p27kip1 function in a normal cell and a cancer cell. In normal cells, p27kip1 is located in the nucleus and arrests G1-S phase progression through the inhibition of the cyclin E-Cdk2 complex. The AKT pathway is inhibited by tumor suppressing mechanisms, such as PTEN. Microtubule formation and degradation is properly controlled by stathmin or other mechanisms. However, in a cancer cell, inactivation of PTEN or the other tumor suppressors lose the control of AKT. p27kip1 is phosphorylated by AKT, which lead to the mislocalization of p27kip1 from nucleus to cytoplasm. Being sequestered into the cytoplasm, p27kip1 is unable to control cyclin E-Cdk2 complex in the nucleus, which permits unchecked cell cycle progression and, therefore, cell proliferation. In addition, cytoplasmic p27kip1 may also affect stathmin, or the other proteins, such as mTOR, which causes increased cell migration.
Report Figure 1A: A majority of human osteosarcoma (OS) cases showed cytoplasmic p27 staining. 1B: p27 was expressed or weakly expressed in the nuclei of the parental non-metastatic OS cells (SaOS-2, DunnOS and K7), but was highly expressed in the cytoplasm of metastatic OS cells (LM7, DLM8 and K7M3).
Figure 2A: p27 was mislocalized in the cytoplasm of the cells transfected with NES-p27 ck-. 2b: Expression of NES-p27 ck- significantly increased the number of migrated cells (stained in purple). 2C: Expression of NES-p27 ck- significantly increased the number of invasive cells (stained in purple).
Figure 3A: Three phosphorylation site mutants (S10A, T157A, and T198A) were generated on NES-p27 ck- using site-directed mutagenesis. 3B: SaOS-2 cells transfected with different phosphorylation site mutants were tested for cell migration. Mutations on Serine-10 (S10A) and Threonine-198 (T198A) significantly decreased the number of migrated cells (stained in purple) relative to NES-p27 ck- .
Report Figure 4: The effect of different concentrations of rapamycin on p27 protein localization in metastatic LM7 cells and its parental non-metastatic SaOS-2 cells by immunohistochemical staining.