Copland, John A.

Abstract Basal subtype breast cancers have a particularly poor prognosis, with high invasiveness and resistance to most targeted therapies. TGFβ and MYC drive central features of basal breast cancer: TGFβ is an autocrine and paracrine signaling factor that drives cell invasion and metastasis, and MYC is a central regulator of cellular proliferation that is upregulated in many cancer types. We show here that genetic or pharmacologic inhibition of MYC in MCF10A basal breast cells results in increased sensitivity to TGFβ-stimulated invasion and metastasis and also show that this signaling loop is dependent on activation of SRC. Analysis of human breast cancer datasets and additional experiments with breast cancer cell lines further suggest the relevance of this signaling loop in basal, but not luminal, breast cancers. Our results imply precaution should be taken when utilizing therapeutic inhibitors of MYC with basal breast cancer patients as this could lead to increased metastasis; however, simultaneous pharmacologic inhibition of SRC and MYC for these patients could facilitate the antiproliferative effects of MYC inhibition while blocking the consequent promotion of metastasis. Cancer Res; 76(12); 3520–30. ©2016 AACR.

Abstract Background Intra-tumor heterogeneity is the driving force behind disease progression and metastasis, drug resistance, and relapse in cancer patients. Defining personalized treatment regimens and increasing the arsenal of targeted therapies available will provide the best chance of prolonging survival and finding curative treatments. Recent work by our group and others has shed light on aberrant de novo lipogenesis as a feature of many cancers including kidney, breast, and lung cancer. Of the molecules involved in fatty acid metabolism, stearoyl-CoA desaturase 1 (SCD1) is frequently over-expressed. Targeted inhibition of SCD1 demonstrates a loss of tumor cell viability in a variety of cancer models. While there currently are specific small molecule inhibitors for SCD1 none are under investigation as anti-cancer therapeutics. In order to evaluate the efficacy of SCD1 inhibitors against ccRCC in both a preclinical and eventually a clinical setting we sought to generate unique compounds and test their biological efficacy. Methods Utilizing an innovative in silico approach, we designed new inhibitors for SCD1 via a scaffold hopping approach while searching for different “core” scaffolds. Critical interaction moieties from the chemical R-groups were held fixed, while generated cores were rapidly scanned for best-fit criteria (shape, docking, pharmacophore fit (QSAR)) into our Z-matrix. Following de novo compound generation and synthesis, we implemented iterative analog generation for the top hits. Using this method and diversity of chemical space, we pursued three divergent species of compounds into the nanomolar regime. Biological testing of the novel compounds included high-throughput proliferative-based screening, oleic acid rescue, and evaluation of the endoplasmic stress response. Enzymatic inhibition of SCD1 was evaluated by LC-MS. Preclinical pharmacokinetics was established using a range of doses administered intravenously and orally. Anti-tumor response was evaluated in an ectopic model of renal carcinoma, both as a single agent and in combination with SOC. Results A cohort of potential unique small molecule inhibitors against SCD1 were designed and synthesized. Of these, a select group of compounds demonstrated potent and selective enzyme inhibition and reduced tumor growth in in vitro and in vivo models of ccRCC. These novel inhibitors reliably reproduced activation of ER stress markers, authenticating this biological response as a harbinger for successful abrogation of SCD1 activity. Conclusions We have established an effective, expedited, and economical format for in silico modeling of novel agents based on a scaffold-hopping methodology. Our findings provide compelling evidence supporting the therapeutic benefit of SCD1 inhibitors for clinical use as broad-spectrum anti-tumor agents alone or in combination with SOC. Citation Format: Christina A. Von Roemeling, Thomas R. Caulfield, Derek C. Radisky, Ilah Bok, Laura A. Marlow, James Miller, Mojda Sidiqi, Anthony B. Pinkerton, Winston W. Tan, Amy L. Lane, Han W. Tun, John A. Copland. Accelerated drug discovery platform yields synthesis of novel stearoyl-CoA desaturase 1 inhibitors that demonstrate anti-tumor efficacy in several models of aggressive cancer. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 4459. doi:10.1158/1538-7445.AM2015-4459

Abstract Acinar cell carcinoma of the pancreas (ACC) is an uncommon malignancy, accounting for less than 1% of all pancreatic neoplasms. Because of its rarity, only a few retrospective studies are available to help guide management. We previously reported the case of a patient with metastatic ACC who achieved prolonged survival with doxorubicin as a result of personalized treatment designed in part on the basis of molecular and in-vitro data collected on analysis of the tumor and primary cells in culture developed from the liver metastasis. We now report the characterization of a patient derived xenograft (PDX) mouse model originating from this patient's ACC liver metastasis tissue. Antitumor activity of multiple drugs (5-FU, irinotecan, oxaliplatin, gemcitabine, bevacizumab, erlotinib, doxorubicin and imatinib) used as single agent therapy is demonstrated. Of the targeted and cytotoxic therapies used, oxaliplatin produced a dramatic and prolonged response to therapy even after withdrawal of treatment. Bevacizumab produced a significant response, as well. Serum lipase and tissue amylase levels correlated with the antitumor response to therapy. Thus, we have developed and characterized an ACC PDX model that may be used in drug discovery for the treatment of this rare cancer for which no standard of care exists. Note: This abstract was not presented at the meeting. Citation Format: Laura A. Marlow, Adam C. Mathias, Louis K. Dawson, William F. Durham, Kenneth A. Meshaw, Robert J. Mullin, Aidan J. Synnott, Ricardo Paz-Fumagalli, Murli Krishna, Daniel Von Hoff, Daniel L. Small, Gerardo Colon-Otero, John A. Copland. Therapeutic responses in a novel patient-derived xenograft mouse model for rare acinar cell pancreatic carcinoma. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 1475. doi:10.1158/1538-7445.AM2015-1475

Abstract Patient derived xenografts (PDX) mouse models for many types of cancer have demonstrated therapeutic responses similar to those seen in patients. Thus, resected patient tumor tissue directly implanted into immune compromised mice followed by therapy and tumor growth analysis is thought to be the closest preclinical model to predict patient therapeutic responses. To date, few models are available for the different histotypes of thyroid cancer derived from follicular thyrocytes; these include papillary, follicular, Hurthle cell, squamous and anaplastic thyroid carcinoma. We have developed eight PDX models in athymic nude mice representing all of these subtypes. The models have been extensively characterized for mutational status (i.e. BRAF, telomerase, RAS, PTEN, TP53, RB, etc) as well as validation by short tandem repeat (STR) analysis to match that of the originating patient tumor tissue. Squamous and anaplastic thyroid cancers are rare tumor types with no FDA approved therapies. Each model demonstrated their own unique responses to radiation, cytotoxic or molecular targeted therapies such as doxorubicin, sorafenib (tyrosine kinase inhibitor), cisplatin, paclitaxel, sunitinib (tyrosine kinase inhibitor), erlotinib (EGFR inhibitor), trametinib (MEK inhibitor) and vemurafenib (BRAF inhibitor). We expect that these models may provide useful in vivo models for thyroid cancer research as well as models for therapeutic guidance based upon histotype, mutational status and response to therapies. Citation Format: Laura A. Marlow, Adam C. Mathias, Louis K. Dawson, William F. Durham, Kenneth A. Meshaw, Robert J. Mullin, Daniel L. Small, Aidan J. Synnott, Dragana Milosevic, Brian C. Netzel, Stefan K. Grebe, Kevin Wu, Robert C. Smallridge, John A. Copland. Characterization of novel thyroid PDX models and their response to combination therapies. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 1458. doi:10.1158/1538-7445.AM2015-1458

Abstract Patient derived xenografts (PDX) mouse models for many types of cancer have been developed recapitulating therapeutic responses in the patient for which the PDX model was derived. Essentially, live surgically resected or biopsy tumor tissues are directly implanted into immune incompetent mice to develop in vivo models to test therapies, follow tumor progression (gene mutations that occur with patient tumor progression do so in mice) or to better understand drug resistance. However, the use of PDX models also has its limitations and challenges. Our laboratory has attempted to develop PDX models in athymic nude mice for numerous cancer histotypes such as bladder, brain, breast, colon, kidney, lung, melanoma, ovarian, pancreas, sarcoma and thyroid. While melanoma and colon cancers have very high tumor take rates (73 - 100%), the other cancer histotypes do not (14 - 50%). We share our experience for each cancer histotype and the techniques used to enhance the success rates in athymic nude mice. We also address the other challenges that include expansion of tumors after cryopreservation and conversion of human tumors to mouse tumors and spontaneous lymphomas. We provide methods to test for human versus mouse content that include QPCR and immunohistochemistry Citation Format: Laura A. Marlow, James L. Miller, Brandy Edenfield, Adam C. Mathias, Louis K. Dawson, William F. Durham, Robert J. Mullin, Daniel L. Small, Aidan J. Synnott, Kevin Wu, John A. Copland. Methods for developing patient derived xenografts in athymic nude mice. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 1453. doi:10.1158/1538-7445.AM2015-1453

Abstract The success of clinical trials for patients with malignant melanoma (MM) designed from the “promising” results using pre-clinical models is minimal. With its increase in incidence, there is a dire need to develop both pre-clinical models comparable to clinical practice that have the capability to translate from bench-to-bedside and novel therapeutic agents for this life-threatening form of skin cancer. To address this concern, we have developed a panel of pre-clinical MM models including patient matched cell lines, patient cell-line derived xenograft models and patient tumor tissue-derived xenograft models. Patient derived xenograft (PDX) models are purported to most closely replicate a patient's response to therapy. Since the development of patient-derived cell lines and tumors grown in mice were more efficacious, their genetic, phenotypic, and therapeutic responses were compared to those of their corresponding patient-matched PDX model. Each model was molecularly characterized to confirm a match to that of the originating patient tumor tissue via protein immunohistochemistry (IHC), mRNA quantitative polymerase chain reaction (qPCR), DNA short tandem repeat (STR) analysis, and Sanger sequencing. Each pre-clinical model provided a genetic match to their respective patient tumor sample. All pre-clinical models were positive for the melanoma-specific markers (S-100 beta, HMB-45, and Melan-A) confirming a melanoma phenotype. Soft agar studies and tumor growth of our PDX and patient cell-line derived xenograft models confirmed tumorigenicity within our models providing the foundation to perform therapeutic studies. Mutation status data (i.e. BRAF, RAS), patient therapeutic clinical history and oncogene overexpression analysis (pERK, pAKT, PI3K, pmTOR, etc.) provide the platform and rationale behind the use of specific therapeutic agents per model. BRAF, NRAS and KRAS mutations were identified along with elevated pERK, pAKT and pmTOR in the models. Targeted therapies (trametinib/MEK inhibitor; vemurafenib/BRAF inhibitor; MK-2206/pAKT inhibitor; and everolimus/mTOR inhibitor) demonstrated dose dependent inhibition of cell proliferation in each of the models based upon oncogenic expression or mutations. In vivo experiments are in progress, which will allow comparison of the three models. The next steps include combination therapy to identify antitumor synergy targeting oncogenic signaling pathways. Citation Format: Antoneicka L. Harris, Laura Marlow, Adam Mathias, Louis Dawson, William Durham, Kenneth Meshaw, Robert Mullin, Daniel Small, Aidan Synnott, Kevin Wu, Dragana Milosevic, Brian Netzel, Stefan Grebe, Svetomir Markovic, John Copland. Characterization of novel patient derived melanoma xenografts and cell lines in response to targeted therapies. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 2417. doi:10.1158/1538-7445.AM2015-2417