Martine Roudier

Sclerostin is expressed in articular cartilage but loss or inhibition does not affect cartilage remodeling during aging or following mechanical injury

OBJECTIVE Sclerostin plays a major role in regulating skeletal bone mass, but its effects in articular cartilage are not known. The purpose of this study was to determine whether genetic loss or pharmacologic inhibition of sclerostin has an impact on knee joint articular cartilage. METHODS Expression of sclerostin was determined in articular cartilage and bone tissue obtained from mice, rats, and human subjects, including patients with knee osteoarthritis (OA). Mice with genetic knockout (KO) of sclerostin and pharmacologic inhibition of sclerostin with a sclerostin-neutralizing monoclonal antibody (Scl-Ab) in aged male rats and ovariectomized (OVX) female rats were used to study the effects of sclerostin on pathologic processes in the knee joint. The rat medial meniscus tear (MMT) model of OA was used to investigate the pharmacologic efficacy of systemic Scl-Ab or intraarticular (IA) delivery of a sclerostin antibody-Fab (Scl-Fab) fragment. RESULTS Sclerostin expression was detected in rodent and human articular chondrocytes. No difference was observed in the magnitude or distribution of sclerostin expression between normal and OA cartilage or bone. Sclerostin-KO mice showed no difference in histopathologic features of the knee joint compared to age-matched wild-type mice. Pharmacologic treatment of intact aged male rats or OVX female rats with Scl-Ab had no effect on morphologic characteristics of the articular cartilage. In the rat MMT model, pharmacologic treatment of animals with either systemic Scl-Ab or IA injection of Scl-Fab had no effect on lesion development or severity. CONCLUSION Genetic absence of sclerostin does not alter the normal development of age-dependent OA in mice, and pharmacologic inhibition of sclerostin with Scl-Ab has no impact on articular cartilage remodeling in rats with posttraumatic OA.

Inhibition of RANKL increases the anti-tumor effect of the EGFR inhibitor panitumumab in a murine model of bone metastasis

Bone metastases cause severe skeletal complications and are associated with osteoclast-mediated bone destruction. RANKL is essential for osteoclast formation, function, and survival, and is the primary effector of tumor-induced osteoclastogenesis and osteolysis. RANKL inhibition by its soluble decoy receptor osteoprotegerin (OPG) prevents tumor-induced osteolysis and decreases skeletal tumor burden. Because osteoclast-mediated bone resorption releases growth factors from the bone matrix, the host bone microenvironment induces a vicious cycle of bone destruction and tumor proliferation and survival. A prediction of this vicious cycle hypothesis is that targeting the host bone microenvironment by osteoclast inhibition would reduce tumor growth and survival and may enhance the anti-tumor effects of targeted therapies. The epidermal growth factor receptor (EGFR) pathway regulates critical processes such as cell growth and survival, and anti-EGFR therapies can cause tumor cell arrest and apoptosis. We evaluated whether reduction of osteolysis by RANKL inhibition could enhance the anti-tumor effects of an anti-EGFR antibody (panitumumab) in a novel murine model of human A431 epidermoid carcinoma bone metastasis. Skeletal tumor progression was assessed longitudinally by bioluminescence imaging. RANKL inhibition by OPG-Fc treatment resulted in a reduction in tumor progression in bony sites. OPG-Fc treatment also caused a dose-dependent reduction in tumor-induced osteolysis, supporting the essential role of RANKL in this process. In combination, RANKL inhibition increased the anti-tumor efficacy of an anti-EGFR antibody, and completely blocked tumor-induced bone breakdown, demonstrating that addition of the indirect anti-tumor effect of RANKL inhibition increases the anti-tumor efficacy of panitumumab, a targeted anti-EGFR antibody.

New variants in the Enpp1 and Ptpn6 genes cause low BMD, crystal-related arthropathy, and vascular calcification.

A large genome‐wide, recessive, N‐ethyl‐N‐nitrosourea (ENU)‐induced mutagenesis screen was performed on a mixed C57BL/6J and C3H.SW‐H2/SnJ mouse background to identify genes regulating bone mass. Approximately 6500 male and female G3 hybrid mice were phenotyped at 8 and 10 wk of age by DXA analysis for evidence of changes in unadjusted or body weight–adjusted BMD or BMC. Phenodeviant lines were identified based on statistical criteria that included a false discovery rate (FDR) <20% and Z‐score >2.8. Genome‐wide mapping scans were initiated on 22 lines, with evidence of high or low BMD or BMC that deviated by approximately −30% to +50% from the means. Several lines were discontinued as showing lack of heritability, but two heritable lines were identified with narrow chromosomal regions that allowed sequencing of potential mutant candidate genes. Novel mutations were identified in the Enpp1 (C397S) gene on chromosome 10 (line 4482) and the Ptpn6 (I482F) gene on chromosome 6 (line 4489) that were both associated with low bone mass. In addition, the phenotype of the Enpp1 mice showed a striking joint disease and calcification of blood vessels including the aorta, myocardium, and renal arteries and capillaries. These results support a role for the Enpp1 gene in the pathogenesis associated with mineralization of articular cartilage and vascular calcification. This work confirms the utility of the chemical mutagenesis approach for identification of potential disease genes and confirms the role of Enpp1 and Ptpn6 in regulating mineralization and skeletal bone mass.

RANKL inhibition combined with tamoxifen treatment increases anti-tumor efficacy and prevents tumor-induced bone destruction in an estrogen receptor-positive breast cancer bone metastasis model

Tumor cells in bone can induce the activation of osteoclasts, which mediate bone resorption and release of growth factors and calcium from the bone matrix, resulting in a cycle of tumor growth and bone breakdown. Targeting the bone microenvironment by the inhibition of RANKL, an essential mediator of osteoclast function, not only prevents tumor-induced osteolysis but also decreases skeletal tumor burden in preclinical models. The inhibition of skeletal tumor progression after the inhibition of osteoclasts is via interruption of the “vicious cycle” of tumor/bone interactions. The majority of breast cancer patients at risk for bone metastases harbor estrogen receptor-positive (ER+) tumors. We developed a mouse model for ER+ breast cancer bone metastasis and evaluated the effect of RANKL inhibition on tumor-induced osteolysis and skeletal tumor growth both alone and in combination with tamoxifen. Luciferase-labeled MCF-7 cells (MCF-7Luc) formed metastatic foci in the hind limbs following intracardiac injection and caused mixed osteolytic/osteoblastic lesions. RANKL inhibition by OPG-Fc treatment blocked osteoclast activity and prevented tumor-induced osteolysis, as well as caused a marked decrease in skeletal tumor burden. Tamoxifen as a single agent reduced MCF-7Luc tumor growth in the hind limbs. In a combination experiment, OPG-Fc plus tamoxifen resulted in significantly greater tumor growth inhibition than either single agent alone. Histologic analysis revealed a decrease in the proliferation of tumor cells by both single agents, which was enhanced in the combination treatment. Upon treatment with OPG-Fc alone or in combination with tamoxifen, there was a complete absence of osteolytic lesions, demonstrating the ability of RANKL inhibition to prevent skeletal related morbidity in an ER+ model. The combination approach of targeting osteoclasts and the bone microenvironment by RANKL inhibition and the tumor directly via hormonal therapy may provide additional benefit to reducing skeletal tumor progression in ER+ breast cancer patients.

Abstract 1583: The biological and molecular characterization of 24 unique prostate cancer xenograft lines, including responses to therapy

Proceedings: AACR 102nd Annual Meeting 2011‐‐ Apr 2‐6, 2011; Orlando, FL Introduction: Prostate cancers (PCa) are heterogeneous which results in an unpredictable and wide range in response to therapy. A significant limitation in unraveling the complexities of PCa and designing/evaluating novel therapeutic strategies is the lack of pre-clinical models that closely replicate the diversity of the disease seen in man. To overcome this limitation we established 24 PCa xenograft lines (LuCaP series). Last year we reported on a subset of these lines. This presentation will provide a detailed biological and molecular characterization of nearly all of these xenograft lines including their responses to therapy. Methods: Xenografts were initiated from tumors acquired at radical prostatectomy or rapid autopsy (enabling acquisition of metastatic foci) and implanted into male immune-compromised mice. Characterization includes: (a) basic histology, (b) serum PSA levels, (c) tumor doubling time, (d) expression of 38 biomarkers by immunohistochemistry (IHC), (e) gene expression array profiles, (f) expression of the androgen receptor (AR) and its splice variants, (g) bone remodeling perturbations associated with tumor growth in bone, and (h) response to therapy, i.e. androgen ablation, docetaxel and anti-IGF-1R. Results: As in the clinical scenario, these xenografts display considerable diversity. All histologically resemble the originating clinical specimen; 20 are adenocarcinomas and 4 are neuroendocrine. Unsupervised clustering revealed (a) association between the xenograft and the originating clinical specimen, (b) pairing of androgen dependent lines with their castrate-resistant offspring, (c) a distinction between adenocarcinoma and neuroendocrine phenotypes and (d) continuous passage in vivo for 2-8 years did not impact gene expression profiles. PSA serum levels range from zero to the low thousands of ng/ml for a 1 g tumor. AR expression is highly variable and >50% of the lines express an AR splice variant after castration. IHC revealed considerable heterogeneity in biomarker expression. In most cases protein expression correlated well with gene expression. The bone remodeling response indicates at least 8 are osteoblastic. Prolonged survival (PS) to therapy was extremely variable; androgen ablation ranged from 1-7.3 fold and docetaxel also from 1-7.3. Interestingly, LuCaP 86.2 expressing predominantly ARv567es was among the least responsive to androgen ablation (PS 1.1) whereas it is one of the most responsive to docetaxel (PS >4). Conclusions: These 24 LuCaP PCa xenograft lines are highly diverse and clinically relevant models for the study of PCa biology. This diversity of phenotypes and responses to therapy most importantly suggests that misleading conclusions can be drawn from use of only one or two PCa models. This is extremely important in the evaluation of new therapeutic strategies, especially those that target specific pathways. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 102nd Annual Meeting of the American Association for Cancer Research; 2011 Apr 2-6; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2011;71(8 Suppl):Abstract nr 1583. doi:10.1158/1538-7445.AM2011-1583

Abstract 3242: The molecular and biological characterization of 25 unique prostate cancer xenograft lines, including response to therapy

Proceedings: AACR 101st Annual Meeting 2010‐‐ Apr 17‐21, 2010; Washington, DC Introduction: Prostate cancers (PCa) exhibit a wide range of molecular and biological profiles. In concert with this heterogeneity, there is a wide spectrum of responses to therapy. One of the significant limitations in unraveling the complexities of PCa and designing/evaluating novel therapeutic strategies has been the lack of pre-clinical models that closely replicate the diversity seen in man. To overcome this limitation we have developed 25 unique PCa xenograft lines (LuCaP series) where the initiating implants were derived from primary tumors, soft-tissue metastases and bone metastases. This presentation will provide details of a comprehensive molecular and biological characterization of these lines, including their responses to therapy. Methods: Tumors were obtained either from radical prostatectomies or from our rapid autopsy program, which is designed specifically for the acquisition of metastatic foci. Tumor pieces were implanted into male, immune compromised mice. Characterization involves: (a) quantification of serum PSA levels, (b) basic histology, (c) tumor doubling time, (d) a panel of >15 biomarkers assessed by immunohistochemistry, (e) gene expression array profiles, (f) chromosomal losses and gains, (g) expression of the TMPRSS2:ERG fusion gene, (h) presence of the androgen receptor (AR) and its splice variants, (i) bone remodeling perturbations associated with tumor growth in bone, and (j) response to therapy, including androgen ablation and docetaxel. Results: Overall, 18% of the attempts to establish xenografts were successful resulting in 25 lines. As in the clinical scenario, these xenografts display a wide range of characterization profiles. While most are adenocarcinomas, 3 are neuroendocrine and all histologically resemble the originating clinical specimen. PSA serum levels can reach into the low thousands of ng/ml for a 1 g tumor. AR gene and protein expression is highly variable and >50% of the lines express a constitutively active AR splice variant. The gene expression profiles show multiple clustering patterns that include a close association of the xenograft with the clinical specimen, pairing of androgen dependent lines with their castrate-resistant offspring and a distinction between adenocarcinoma and neuroendocrine histologies. Five of the lines are PTEN negative and 9 express the TMPRSS2:ERG gene fusion. The bone remodeling response ranges from osteoblastic to lytic. Responses to androgen ablation and docetaxel range from long duration to no response. Conclusions: These 25 LuCaP PCa xenograft lines provide a highly diverse, yet clinically relevant, panel for the study of PCa biology. This diversity most importantly suggests that misleading conclusions can be drawn from use of only one or two PCa models. This is extremely important in the evaluation of new therapeutic strategies, especially those that target specific pathways. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 101st Annual Meeting of the American Association for Cancer Research; 2010 Apr 17-21; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2010;70(8 Suppl):Abstract nr 3242.

Abstract 3302: Profiling prostate cancer metastases

Introduction: Prostate cancer (PCa) metastasizes to a number of sites, however, it has a propensity to metastasize to bone. There is considerable variation in sites of metastases, the phenotype of metastases, and bone response in PCa. Our objective was to detail the behavior of PCa metastases in a large cohort of patients who have died of PCa. Methods: Tissue samples used in this study were obtained from seventy patients who had died from advanced PCa and who underwent a rapid autopsy, which was performed under the aegis of the Prostate Cancer Donor Program at the University of Washington. We evaluated the clinical history, sites of metastasis, pathology, biomarker expression and bone response for all 70 patients. We used laser capture microdissection to isolate mRNA from liver, lymph node and bone metastases and performed expression analyses on Agilent TM Oligo Arrays (n=32). Tissue microarrays were used to determine the number of neuroendocrine metastases, androgen receptor positivity, androgen-associated protein expression, and PCa proliferation rates (Ki-67) from 42 patients. Histomorphometry was used to evaluate bone response in ∼20 predetermined bone biopsy sites from 50 of the patients. Results: PCa bone metastases occurred in 90% of our patients that died of PCa. Additionally, the majority of bone metastases were predominantly osteoblastic 71%, with 20% predominantly osteolytic and 9% with a mixed or no bone response. The predominant non-bone sites were lymph nodes, liver, lung and adrenal glands respectively. Seven of 42 patients had metastases with a neuroendocrine phenotype. Nuclear androgen receptor expression was high in bone and low in liver metastases and the correlation between nuclear AR and cytoplasmic PSA was 0.49. Proliferation rates were low in normal prostate increasing in Gleason 3 through 4, and higher in metastases. In a preliminary expression analysis we observed significantly higher expression of CD302, ZNF329, and TMEM79 in the PCa bone metastases compared to soft tissue metastases. Interestingly, bone histomorphometry revealed that the number of TRAcP positive osteoclasts was halved in bone cores from patients treated with bisphosphonates (n=459) compared to bone cores from non-bisphosphonate treated patients (n=284) (p=0.0003). Conclusions: The heterogeneity of PCa metastases both within and between patients is a considerable obstacle in determining the appropriate treatment regime for each individual. Our data highlight these differences, while suggesting patients may be grouped into cohorts with consistent molecular signatures and clinical outcomes. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 101st Annual Meeting of the American Association for Cancer Research; 2010 Apr 17-21; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2010;70(8 Suppl):Abstract nr 3302.