A major obstacle in developing effective therapies against solid tumors stems from an inability to adequately model the rare subpopulation of panresistant cancer cells that may often drive the disease. We describe a strategy for optimally modeling highly abnormal and highly adaptable human triple-negative breast cancer cells, and evaluating therapies for their ability to eradicate such cells. To overcome the shortcomings often associated with cell culture models, we incorporated several features in our model including a selection of highly adaptable cancer cells based on their ability to survive a metabolic challenge. We have previously shown that metabolically adaptable cancer cells efficiently metastasize to multiple organs in nude mice. Here we show that the cancer cells modeled in our system feature an embryo-like gene expression and amplification of the fat mass and obesity associated gene FTO. We also provide evidence of upregulation of ZEB1 and downregulation of GRHL2 indicating increased epithelial to mesenchymal transition in metabolically adaptable cancer cells. Our results obtained with a variety of anticancer agents support the validity of the model of realistic panresistance and suggest that it could be used for developing anticancer agents that would overcome panresistance.

We have previously shown that only 0.01% cells survive a metabolic challenge involving lack of glutamine in culture medium of SUM149 triple-negative Inflammatory Breast Cancer cell line. These cells, designated as SUM149-MA for metabolic adaptability, are resistant to chemotherapeutic drugs, and they efficiently metastasize to multiple organs in nude mice. We hypothesized that obesity-related molecular networks, which normally help in cellular and organismal survival under metabolic challenges, may help in the survival of MA cells. The fat mass and obesity-associated protein FTO is overexpressed in MA cells. Obesity-associated cis-acting elements in non-coding region of FTO regulate the expression of IRX3 gene, thus activating obesity networks. Here we found that IRX3 protein is significantly overexpressed in MA cells (5 to 6-fold) as compared to the parental SUM149 cell line, supporting our hypothesis. We also obtained evidence that additional key regulators of energy balance such as ARID5B, IRX5, and CUX1 P200 repressor could potentially help progenitor-like TNBC cells survive in glutamine-free medium. MO-I-500, a pharmacological inhibitor of FTO, significantly (>90%) inhibited survival and/or colony formation of SUM149-MA cells as compared to untreated cells or those treated with a control compound MO-I-100. Curiously, MO-I-500 treatment also led to decreased levels of FTO and IRX3 proteins in the SUM149 cells initially surviving in glutamine-free medium as compared to MO-I-100 treatment. Interestingly, MO-I-500 treatment had a relatively little effect on cell growth of either the SUM149 or SUM149-MA cell line when added to a complete medium containing glutamine that does not pose a metabolic challenge. Importantly, once selected and cultured in glutamine-free medium, SUM149-MA cells were no longer affected by MO-I-500 even in Gln-free medium. We conclude that panresistant MA cells contain interconnected molecular networks that govern developmental status and energy balance, and genetic and epigenetic alterations that are selected during cancer evolution.

We treated highly metabolically adaptable (SUM149-MA) triple-negative inflammatory breast cancer cells and their control parental SUM149-Luc cell line with JQ1 for long periods to determine its efficacy at inhibiting therapy-resistant cells. After 20 days of treatment with 1–2 µM of JQ1, which killed majority of cells in the parental cell line, a large number of SUM149-MA cells survived, consistent with their pan-resistant nature. Interestingly, though, the JQ1 treatment sensitized resistant cancer cells in both the SUM149-MA and SUM149-Luc cell lines to subsequent treatment with doxorubicin and paclitaxel. To measure JQ1-mediated sensitization of resistant cancer cells, we first eradicated approximately 99% of relatively chemotherapy-sensitive cancer cells in culture dishes by long treatments with doxorubicin or paclitaxel, and then analyzed the remaining resistant cells for survival and growth into colonies. In addition, combination, rather than sequential, treatment with JQ1 and doxorubicin was also effective in overcoming resistance. Notably, Western blotting showed that JQ1-treated cancer cells had significantly lower levels of PD-L1 protein than did untreated cells, indicating that JQ1 treatment may reduce tumor-mediated immune suppression and improve the response to immunotherapy targeting PD-L1. Finally, JQ1 treatment with a low 62.5 nM dose sensitized another resistant cell line, FC-IBC02-MA, to treatment with doxorubicin and paclitaxel.

Background: Cell culture models of cancer typically favor proliferative but therapy-sensitive cells because body-like selection pressures are absent. To address this limitation of cell culture, we previously described a function-based selection strategy to model deep intrinsic resistance in cultures of triple-negative breast cancer cells. To determine the validity of this approach in identifying noncytotoxic drugs that could inhibit the relapse of poor-prognosis minimal residual disease in breast cancer, we used our novel cell culture model to evaluate a well-known BET bromodomain inhibitor, JQ1, which modulates cancer epigenome. Methods: We treated highly metabolically adaptable (SUM149-MA) cells and their control parental SUM149-Luc cell line with JQ1 for long periods to determine its efficacy at inhibiting resistant cells. To measure JQ1-mediated sensitization of resistant cancer cells, we first eradicated approximately 99% of relatively chemotherapy-sensitive cancer cells in culture dishes by long treatments with doxorubicin or paclitaxel, and then analyzed the remaining resistant cells for survival and growth into colonies. Our methods were designed to reveal resistant cancer cells while minimizing the contribution of the vast majority of nonresistant cells that simply proliferate in cell culture. We also performed Western blotting to determine whether JQ1 treatment affected PD-L1 protein levels in cancer cells. Results: After 20 days of treatment with 1-2 µM JQ1, which killed a majority of cells in the parental cell line, a large number of SUM149-MA cells survived, consistent with their pan-resistant nature. Interestingly, though, the JQ1 treatment sensitized resistant cancer cells in both the SUM149-MA and SUM149-Luc cell lines to subsequent treatment with doxorubicin and paclitaxel. In addition, combination, rather than sequential, treatment with JQ1 and doxorubicin was also effective in overcoming resistance. Notably, JQ1-treated cancer cells also had significantly lower levels of PD-L1 protein than did untreated cells. Conclusions: Our results suggest that the noncytotoxic drug JQ1 could inhibit the growth of resistant breast cancer cells at the minimal residual disease stage, prior to relapse. Because it can also reduce PD-L1 protein levels, JQ1 may reduce tumor-mediated immune suppression and improve response to immune therapy targeting PD-L1. In all, our results support the validity of a cell culture-based approach for modeling a cancer adaptability phenotype, namely the opportunistic switching of cancer cells between quiescence and proliferation.

A major obstacle in developing effective therapies against solid tumors stems from an inability to adequately model the rare subpopulation of panresistant cancer cells that may often drive the disease. We propose a strategy for optimally modeling highly abnormal and highly adaptable human triple-negative breast cancer (TNBC) cells, and evaluating therapies for their ability to eradicate such cells. To overcome the shortcomings often associated with cell culture models, we have incorporated several features in our model including a selection of highly adaptable cancer cells based on their ability to survive a metabolic challenge. We have shown that metabolically adaptable (MA) TNBC cells efficiently metastasize to multiple organs in nude mice. Recently we have found that MA cells selected from SUM149 cell line feature an embryo-like gene expression, which may be important in generating tumor heterogeneity and therapy-resistance. Our results obtained thus far with a variety of anticancer agents support the validity of the model of realistic panresistance and suggest that it could be used for developing anticancer agents that would overcome panresistance. The cancer cells most responsible for panresistance are heterogeneous, and the likelihood that a particular test therapy will affect these cells is difficult to predict. Our testing strategy, which involves long-term evaluation of anticancer agents on MA cells in parallel with the parental cell line, is optimized to investigate the effects of potential therapeutic compounds on panresistance. This approach can not only help choose superior therapies for clinical trials, but also suggest strategies for overcoming therapeutic resistance in a timely manner. Considering an important role of chromatin in panresistance, we evaluated two promising chromatin modifiers- a bromodomain inhibitor JQ1 and a Jumonji histone demethylase inhibitor JIB-04 for their ability to overcome panresistance in our model. Both these compound appear promising in different cancer models. We are evaluating these compounds for efficacy in eradicating MA cells; we are testing these compounds to determine whether they are effective as single-agents and/or whether they can sensitize MA cells to chemotherapies. Our results indicate that a long treatment with either JQ1 or JIB-04 sensitizes MA cells to chemotherapeutic drugs doxorubicin and paclitaxel. Our results support the utility of our approach in discovering effective anticancer therapies against a heterogeneous and evolving disease like TNBC. Supported by a State of Texas Grant for Rare and Aggressive Cancers. Citation Format: Balraj Singh, Ryan D. Milligan, Hannah E. Kinne, Anna Shamsnia, Anthony Lucci. Highly adaptable triple-negative breast cancer cells as a suitable model for testing epigenetic 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 1719. doi:10.1158/1538-7445.AM2015-1719