Graduation Year


Document Type




Degree Name

Doctor of Philosophy (Ph.D.)

Degree Granting Department

Biology (Integrative Biology)

Major Professor

Conor C. Lynch, Ph.D.

Committee Member

Brian Ruffell, Ph.D.

Committee Member

Srikumar Chellappan, Ph.D.

Committee Member

Eric K. Lau, Ph.D.

Committee Member

Claire Edwards, Ph.D.


Computational Model, Matrix Metalloproteinase, Multiple Myeloma, Prostate Cancer, Bone Remodeling, Bone Metastasis


Bone is a common site of metastasis for many solid malignancies. Bone-metastatic cancers pose a significant clinical problem worldwide and is among the main causes for cancer patient morbidity and mortality. Patients with advanced bone-metastatic diseases often present with either osteolytic or osteogenic bone diseases as their cancers progress. These bone pathologies are products of the cancer co-opting the local bone remodeling stroma to yield important growth nutrients and factors. Unfortunately, skeletal metastases remain incurable and are fatal. Identifying and understanding the causal multicellular and molecular interactions underlying skeletal malignancies can yield crucial ideas for targeting and inhibiting disease progression. In this thesis, we focused on two major skeletal malignancies: multiple myeloma and prostate cancer.

Multiple myeloma frequently induces pathologic bone loss in patients by enhancing osteoclastic bone resorption. Matrix metalloproteinase 13 (MMP-13) is highly abundant in the bone and is primarily expressed by mesenchymal stromal cells (MSCs) in multiple myeloma. Analysis of public datasets and patient biopsies reveal multiple myeloma induce upregulation of MMP-13 in MSCs. Myeloma-induced MMP-13 processes and activates cytokines secreted by MSCs which are important for osteoclast formation and activity, such as CXCL7. Depletion of MMP-13 and/or CXCL7 in MSCs significantly impede MSC induction of osteoclastogenesis. Genetic and pharmacologic ablation of MMP-13 significantly improves overall survival myeloma mouse models, providing clear rationale for the clinical translation of MMP-13 inhibitors for the treatment of multiple myeloma and osteolysis.

Our second area of interest revolves around prostate cancer. Primary prostate cancer is largely curable. Patients diagnosed with disseminated disease often undergo periods of remission owing to chemotherapies and androgen deprivation therapies, but a significant subset eventually relapse with castrate-resistant prostate cancer (CRPC). CRPCs frequently manifests in the bone and establish mixed bone pathologies. Androgen receptor variant 7 expression is a common mechanism of castrate resistance and induces the expression of androgen vasopressin 1A (AVPR1A), an important regulator of the mitogenic pathway. We demonstrated that AVPR1A is vital for growth and maintenance of drug-resistance in CRPC cells. Pharmacologic inhibition of AVPR1A using Relcovaptan significantly reduced CRPC tumor growth and improved overall survival in a subcutaneous, an orthotopic, and a bone metastasis mouse model. Specifically, in the bone metastasis model, AVPR1A is expressed by stromal osteoclasts and regulates bone resorption. Relcovaptan-treated tumor-bearing mice also demonstrated significantly reduced cancer-induced bone pathology, supporting the clinical translation of Relcovaptan for the treatment of advanced stage prostate cancer.

Our understanding of molecular and cellular drivers of tumor-bone interactions continue to expand. Dysregulated bone remodeling lies in the center of various skeletal malignancies and involves various bone marrow cell types in addition to osteoclasts, osteoclasts and cancer cells. An emerging question is, “how do all of these components interact with each other over time in cancer-induced osteal pathologies?” To address this, we utilized mathematical models and here we present our findings on using unique modeling approaches to define cellular interactions in the bone microenvironment. Mathematical modeling allows for describing the complex relationships between various cellular species, and predicting the consequences of disrupting these relationships. To eventually build models of skeletal malignancies as prostate cancer and multiple myeloma, we began by building a model of normal bone healing response to injury. We devised an ordinary differential equation (ODE)-based model that is powered by in vivo experimental data to ensure biologically-relevant simulations. Our bone injury model was able to discover specific osteoblast and osteoclast behaviors that are currently difficult to test experimentally, but it was also able to accurately predict temporal dynamics of bone cell population data in an independent biological study. These findings demonstrate that we have a computational tool at our disposal to begin modeling skeletal malignancies and interrogating various treatment options, with hopes to inform clinical translational efforts.

Taken together, our studies herein address important cellular and molecular factors that govern cancer-bone interactions and cancer-induced bone disease, and reveal a number of therapeutic opportunities for the treatment of these incurable lesions.