Tumor-immune metaphenotypes orchestrate an evolutionary bottleneck that promotes metabolic transformation

Introduction

Metabolism plays a complex role in the evolution of cancerous tumors, including inducing a multifaceted effect on the immune system to aid immune escape. Immune escape is, by definition, a collective phenomenon by requiring the presence of two cell types interacting in close proximity: tumor and immune. The microenvironmental context of these interactions is influenced by the dynamic process of blood vessel growth and remodelling, creating heterogeneous patches of well-vascularized tumor or acidic niches.

Methods

Here, we present a multiscale mathematical model that captures the phenotypic, vascular, microenvironmental, and spatial heterogeneity which shapes acid-mediated invasion and immune escape over a biologically-realistic time scale. The model explores several immune escape mechanisms such as i) acid inactivation of immune cells, ii) competition for glucose, and iii) inhibitory immune checkpoint receptor expression (PD-L1). We also explore the efficacy of anti-PD-L1 and sodium bicarbonate buffer agents for treatment. To aid in understanding immune escape as a collective cellular phenomenon, we define immune escape in the context of six collective phenotypes (termed "meta-phenotypes"): Self-Acidify, Mooch Acid, PD-L1 Attack, Mooch PD-L1, Proliferate Fast, and Starve Glucose.

Results

Fomenting a stronger immune response leads to initial benefits (additional cytotoxicity), but this advantage is offset by increased cell turnover that leads to accelerated evolution and the emergence of aggressive phenotypes. This creates a bimodal therapy landscape: either the immune system should be maximized for complete cure, or kept in check to avoid rapid evolution of invasive cells. These constraints are dependent on heterogeneity in vascular context, microenvironmental acidification, and the strength of immune response.

Discussion

This model helps to untangle the key constraints on evolutionary costs and benefits of three key phenotypic axes on tumor invasion and treatment: acid-resistance, glycolysis, and PD-L1 expression. The benefits of concomitant anti-PD-L1 and buffer treatments is a promising treatment strategy to limit the adverse effects of immune escape.

Modeling phenotypic heterogeneity towards evolutionarily inspired osteosarcoma therapy

Osteosarcoma is the most common bone sarcoma in children and young adults. While universally delivered, chemotherapy only benefits roughly half of patients with localized disease. Increasingly, intratumoral heterogeneity is recognized as a source of therapeutic resistance. In this study, we develop and evaluate an in vitro model of osteosarcoma heterogeneity based on phenotype and genotype. Cancer cell populations vary in their environment-specific growth rates and in their sensitivity to chemotherapy. We present the genotypic and phenotypic characterization of an osteosarcoma cell line panel with a focus on co-cultures of the most phenotypically divergent cell lines, 143B and SAOS2. Modest environmental (pH, glutamine) or chemical perturbations dramatically shift the success and composition of cell lines. We demonstrate that in nutrient rich culture conditions 143B outcompetes SAOS2. But, under nutrient deprivation or conventional chemotherapy, SAOS2 growth can be favored in spheroids. Importantly, when the simplest heterogeneity state is evaluated, a two-cell line coculture, perturbations that affect the faster growing cell line have only a modest effect on final spheroid size. Thus the only evaluated therapies to eliminate the spheroids were by switching therapies from a first strike to a second strike. This extensively characterized, widely available system, can be modeled and scaled to allow for improved strategies to anticipate resistance in osteosarcoma due to heterogeneity.

Treatment of evolving cancers will require dynamic decision support

Cancer research has traditionally focused on developing new agents, but an underexplored question is that of the dose and frequency of existing drugs. Based on the modus operandi established in the early days of chemotherapies, most drugs are administered according to predetermined schedules that seek to deliver the maximum tolerated dose and are only adjusted for toxicity. However, we believe that the complex, evolving nature of cancer requires a more dynamic and personalized approach. Chronicling the milestones of the field, we show that the impact of schedule choice crucially depends on processes driving treatment response and failure. As such, cancer heterogeneity and evolution dictate that a one-size-fits-all solution is unlikely-instead, each patient should be mapped to the strategy that best matches their current disease characteristics and treatment objectives (i.e. their 'tumorscape'). To achieve this level of personalization, we need mathematical modeling. In this perspective, we propose a five-step 'Adaptive Dosing Adjusted for Personalized Tumorscapes (ADAPT)' paradigm to integrate data and understanding across scales and derive dynamic and personalized schedules. We conclude with promising examples of model-guided schedule personalization and a call to action to address key outstanding challenges surrounding data collection, model development, and integration.

Intermetastatic and Intrametastatic Heterogeneity Shapes Adaptive Therapy Cycling Dynamics

Adaptive therapies that alternate between drug applications and drug-free vacations can exploit competition between sensitive and resistant cells to maximize the time to progression. However, optimal dosing schedules depend on the properties of metastases, which are often not directly measurable in clinical practice. Here, we proposed a framework for estimating features of metastases through tumor response dynamics during the first adaptive therapy treatment cycle. Longitudinal prostate-specific antigen (PSA) levels in 16 patients with metastatic castration-resistant prostate cancer undergoing adaptive androgen deprivation treatment were analyzed to investigate relationships between cycle dynamics and clinical variables such as Gleason score, the change in the number of metastases over a cycle, and the total number of cycles over the course of treatment. The first cycle of adaptive therapy, which consists of a response period (applying therapy until 50% PSA reduction), and a regrowth period (removing treatment until reaching initial PSA levels), delineated several features of the computational metastatic system: larger metastases had longer cycles; a higher proportion of drug-resistant cells slowed the cycles; and a faster cell turnover rate sped up drug response time and slowed regrowth time. The number of metastases did not affect cycle times, as response dynamics were dominated by the largest tumors rather than the aggregate. In addition, systems with higher intermetastasis heterogeneity responded better to continuous therapy and correlated with dynamics from patients with high or low Gleason scores. Conversely, systems with higher intrametastasis heterogeneity responded better to adaptive therapy and correlated with dynamics from patients with intermediate Gleason scores.

SIGNIFICANCE

Multiscale mathematical modeling combined with biomarker dynamics during adaptive therapy helps identify underlying features of metastatic cancer to inform treatment decisions.

Virtual alignment of pathology image series for multi-gigapixel whole slide images

Interest in spatial omics is on the rise, but generation of highly multiplexed images remains challenging, due to cost, expertise, methodical constraints, and access to technology. An alternative approach is to register collections of whole slide images (WSI), generating spatially aligned datasets. WSI registration is a two-part problem, the first being the alignment itself and the second the application of transformations to huge multi-gigapixel images. To address both challenges, we developed Virtual Alignment of pathoLogy Image Series (VALIS), software which enables generation of highly multiplexed images by aligning any number of brightfield and/or immunofluorescent WSI, the results of which can be saved in the ome.tiff format. Benchmarking using publicly available datasets indicates VALIS provides state-of-the-art accuracy in WSI registration and 3D reconstruction. Leveraging existing open-source software tools, VALIS is written in Python, providing a free, fast, scalable, robust, and easy-to-use pipeline for registering multi-gigapixel WSI, facilitating downstream spatial analyses.

Abstract 5694: Adaptive treatment scheduling of PARP inhibitors in ovarian cancer: Using mathematical modeling to assess clinical feasibility and estimate potential benefits

PARP inhibitors (PARPis) are revolutionizing the treatment of ovarian cancer. Yet for many patients these improvements come at the cost of physical and financial toxicity and responses are typically temporary due to emerging drug resistance. A growing body of pre-clinical and clinical work suggests that when cure is unlikely, it is possible to delay progression and reduce drug use through patient-specific drug scheduling. So-called 'adaptive therapy' dynamically adjusts treatment to preserve drug-sensitive cells which interfere with resistant cells through competition. In a prior study, we developed a mathematical model to describe the treatment response of ovarian cancer cells to the PARPi olaparib in vitro, and we proposed a candidate adaptive PARPi algorithm. Here, we extend our model to capture the dynamics in patients and use it to study the feasibility and potential benefit of adaptive PARPi administration in practice.

Our prior model posited that treatment induces cell cycle arrest that moves cells from the proliferating subset of the population to an arrested compartment. The model predicted that while there is scope for treatment reductions, these need to be carefully timed and prolonged treatment breaks should be avoided. Based on these results we proposed an adaptive treatment algorithm in which the olaparib dose is switched between high and low doses, depending on the tumor’s growth rate. To test the translational potential of this strategy, we retrospectively collected data from 53 ovarian cancer patients who received olaparib at the H Lee Moffitt Cancer Center between 2014 and 2021. Using serum CA-125 as a proxy for tumor burden, we first examined whether our mathematical model could capture the patients’ longitudinal dynamics. While the response of some patients was consistent with what we had observed in vitro, in others there was evidence of the emergence of a distinct drug-resistant population, and we extended our mathematical model to account for this. After calibrating and validating the model with the patient data, we tested whether these patients would have benefited from adaptive PARPi treatment. Our simulations suggest that our proposed algorithm is feasible and provides a means for reducing treatment in a patient-specific manner. In addition, in a subset of patients we predict that adaptive therapy could delay progression. Overall, this work corroborates the potential for adaptive PARPi therapy and helps to identify outstanding challenges on the way to clinical translation.

Citation Format: Maximilian A. Strobl, Alexandra L. Martin, Christopher Gallagher, Mehdi Damaghi, Mark Robertson-Tessi, Robert Gatenby, Robert M. Wenham, Philip K. Maini, Alexander R. Anderson. Adaptive treatment scheduling of PARP inhibitors in ovarian cancer: Using mathematical modeling to assess clinical feasibility and estimate potential benefits. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 5694.

Abstract 847: Applying the principles of convexity and concavity to guide treatment scheduling of ALK inhibitors in non-small cell lung cancer

Recently, the US Food and Drug Administration (FDA) has initiated “Project Optimus,” designed to revisit and reform the decades-old maximum-tolerable dose paradigm. Traditional dose selection schemes designed for cytotoxic therapies may not apply for targeted therapies with exposure-response curves that often plateau below maximum tolerable patient toxicity levels. While nonlinear sigmoidal exposure-response curves are ubiquitous in oncology, they are typically used to measure differential response in first-order effects (mean value of drug dose delivered), while second-order effects (variance of drug dose) are generally ignored. Herein we present a paradigm for guiding treatment scheduling based on the convexity (or its converse, concavity) of exposure-response curves - a phenomenon that is widely applicable to targeted therapies. The “curvature” of exposure-response curves (e.g. a convex or concave shape) provides a direct prediction of continuous treatment, compared with high-dose/low-dose treatment. For example, if the exposure-response function is concave near a dose of ‘x’, continuous (daily) administration of x may be less effective response compared to a regimen that switches equally between 120% of x and 80% of x (every other day), even though both regimens use the same total drug. This paradigm is specifically relevant to targeted therapies which, as stated above, are associated with exposure-response plateaus that are concave in nature (to be compared with convex exposure-response curves common in cytotoxic chemotherapies). Analysis of response curves in vitro for a H3122 ALK-positive non-small cell lung cancer (NSCLC) cell line predicts that evolved-resistance lines are generally more concave, while treatment-naïve lines are more convex. In vivo, selection pressure due to treatment selects for resistant phenotypes over time. Previous literature shows resistance to ALK inhibitors occurs gradually, as tumors acquire cooperating genetic and epigenetic adaptive changes. Thus, we hypothesized the existence of a critical point in the time-evolution of ALK-positive tumors where it is optimal to switch from continuous treatment to high-dose/low-dose to mitigate the onset of gradual resistance. In this work, we construct a mathematical modeling framework of gradual resistance, parameterized to data, and predict time-dependent curvature in continuous (8 weeks), volatile (8 weeks) ALK inhibition in vivo. our key insight is that curvature increases in proportion to the amount of resistance in the tumor population. Curvature provides a time-dependent metric which 1) predicts the emergence of resistance and 2) determines the optimal subsequent dosing strategy. We test our key hypothesis in vivo, comparing continuous and volatile treatment schedules of ALK inhibitors to a switching schedule of continuous-volatile (4 weeks each).

Citation Format: Jeffrey West, Bina Desai, Maximilian Strobl, Luke Pierik, Robert Vander Velde, Mark Robertson-Tessi, Andriy Marusyk, Alexander R. Anderson. Applying the principles of convexity and concavity to guide treatment scheduling of ALK inhibitors in non-small cell lung cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 847.

Agent-based methods facilitate integrative science in cancer

In this opinion, we highlight agent-based modeling as a key tool for exploration of cell-cell and cell-environment interactions that drive cancer progression, therapeutic resistance, and metastasis. These biological phenomena are particularly suited to be captured at the cell-scale resolution possible only within agent-based or individual-based mathematical models. These modeling approaches complement experimental work (in vitro and in vivo systems) through parameterization and data extrapolation but also feed forward to drive new experiments that test model-generated predictions.

Adaptive therapy for ovarian cancer: An integrated approach to PARP inhibitor scheduling

Toxicity and emerging drug resistance are important challenges in PARP inhibitor (PARPi) treatment of ovarian cancer. Recent research has shown that evolutionary-inspired treatment algorithms which adapt treatment to the tumor's treatment response (adaptive therapy) can help to mitigate both. Here, we present a first step in developing an adaptive therapy protocol for PARPi treatment by combining mathematical modelling and wet-lab experiments to characterize the cell population dynamics under different PARPi schedules. Using data from in vitro Incucyte Zoom time-lapse microscopy experiments and a step-wise model selection process we derive a calibrated and validated ordinary differential equation model, which we then use to test different plausible adaptive treatment schedules. Our model can accurately predict the in vitro treatment dynamics, even to new schedules, and suggests that treatment modifications need to be carefully timed, or one risks losing control over tumour growth, even in the absence of any resistance. This is because our model predicts that multiple rounds of cell division are required for cells to acquire sufficient DNA damage to induce apoptosis. As a result, adaptive therapy algorithms that modulate treatment but never completely withdraw it are predicted to perform better in this setting than strategies based on treatment interruptions. Pilot experiments in vivo confirm this conclusion. Overall, this study contributes to a better understanding of the impact of scheduling on treatment outcome for PARPis and showcases some of the challenges involved in developing adaptive therapies for new treatment settings.

Mistic: An open-source multiplexed image t-SNE viewer

Understanding the complex ecology of a tumor tissue and the spatiotemporal relationships between its cellular and microenvironment components is becoming a key component of translational research, especially in immuno-oncology. The generation and analysis of multiplexed images from patient samples is of paramount importance to facilitate this understanding. Here, we present Mistic, an open-source multiplexed image t-SNE viewer that enables the simultaneous viewing of multiple 2D images rendered using multiple layout options to provide an overall visual preview of the entire dataset. In particular, the positions of the images can be t-SNE or UMAP coordinates. This grouped view of all images allows an exploratory understanding of the specific expression pattern of a given biomarker or collection of biomarkers across all images, helps to identify images expressing a particular phenotype, and can help select images for subsequent downstream analysis. Currently, there is no freely available tool to generate such image t-SNEs.

Abstract 5037: Distinct tumor-immune ecologies in NSCLC patients predict progression and define a clinical biomarker of therapy response

Rationale: Examination of multiplexed images of tissues has recently emerged as a routine clinical procedure for cancer diagnosis and prognosis. The simultaneous detection of numerous biomarkers enables the interpretation of cellular states and the characterization of tumor-immune interactions in situ and at the single-cell level. However, image processing and the subsequent interpretive and predictive tools for multiplexed image data remain limited.

Methods: We developed a computational multiplexed-image analysis pipeline using cell-segmentation and quadrat-based approaches to analyze the spatial and temporal features of multiplexed non-small cell lung cancer (NSCLC) images, and predict disease progression and identify clinical biomarkers. Images were obtained from nine patients with advanced/metastatic NSCLC who were treated with the oral HDAC inhibitor vorinostat combined with the PD-1 inhibitor pembrolizumab. Images were collected from all patients both pre- and on-treatment (during the third week).

Results: Both cell-segmentation and quadrat-based approaches confirm that different spatial neighborhoods exist that distinguish progressors (PD) from non-progressors (SD): PD patients have distinct ecologies with higher colocalization of PanCK+PD-1+FoxP3 indicating an immunosuppressive environment, whereas SD patients have a higher colocalization of PanCK+PD-L1 along with T cells suggesting immunoactive tumor regions. These can be considered as potential biomarker candidates for predicting tumor progression. Further, from the single-cell analysis, we note there is a higher abundance of immune cells across the tumor border in PD patients than SD patients. Using the quadrat approach for species distribution modeling, we were able to predict treatment response with 91.4 percent accuracy given each patient’s spatial distribution of cell types from pre-treatment images. Further, we can generate risk maps for each image to identify tumor areas indicating higher probabilities of progression during treatment.

Conclusions: We leveraged both single-cell and quadrat-resolution analysis of multiplexed imaging data and identified fundamentally distinct spatial ecologies between PD and SD patients. The ecology in PD patients appears to be primed for immune resistance even before treatment. This ecological diversity between SD and PD patients acts as a biomarker that enables accurate disease progression prediction.

Citation Format: Sandhya Prabhakaran, Chandler Gatenbee, Mark Robertson-Tessi, Amer A. Beg, Jhanelle Gray, Scott Antonia, Robert A. Gatenby, Alexander R. Anderson. Distinct tumor-immune ecologies in NSCLC patients predict progression and define a clinical biomarker of therapy response [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 5037.

Abstract PR010: Evolutionary Tumor Board: Implementing dynamic personalized therapy using evolutionary theory and mathematical modeling for clinical decision support

The era of big data in oncology has led to the promise of precision medicine for individual patients. However, many therapy decisions continue to be made based on “one size fits most” approaches, primarily since there exist few theoretical and practical tools to deal with a patient’s data over time. In parallel with this growing interest in personalized medicine, cancer is being increasingly recognized as an eco-evolutionary system that adapts to treatments, suggesting that static therapy regimens are often doomed to eventual failure. Here, we present preliminary results from a novel pilot clinical trial (NCT04343365), the Evolutionary Tumor Board (ETB), which uses eco-evolutionary theory (based on experiments and modeling) to assist with clinical decision making for each patient. We developed an informational and computational framework for applying evolutionary therapy approaches to individual patients in a dynamic fashion, using their clinical data in real time. The framework relies on detailed data curation and imaging measurements for each patient, as well as a mathematical modeling approach that accounts for multi-lesion tumor growth, treatment-induced death, and the evolution of resistance. The models are calibrated by historical datasets of similar patients, as well as the patient’s own temporal data. We use a “Phase i trial” approach to account for prediction uncertainty and provide decision support for therapy options available to the patient at any given time point. Crucially, this is presented in a way that harmonizes with the treating oncologist’s intuition. Fifteen patients at Moffitt have been enrolled into the ETB, many of whom have proceeded through the entire process, including follow-up analysis. The ETB generated outcome predictions and therapy recommendations for each case, and subsequent follow-up predictions and recommendations. Our current results demonstrate that the ETB approach has provided both novel and useful decision support for the clinicians. At the same time, numerous opportunities for further research and development have been identified. Our efforts show that there are both challenges and opportunities in the area of personalized therapy, particularly in the context of real-time clinical care. Early results from the ETB show great promise for improving patient outcomes in cancer using mathematical modeling and evolutionary therapy.

Citation Format: Mark Robertson-Tessi, Joel Brown, Maria Poole, Kimberly Luddy, Andriy Marusyk, Jill Gallaher, Jeffrey West, Matthew Johnson, Heiko Enderling, Rikesh Makanji, Joaquim Farinhas, Robert Gatenby, Damon Reed, Christine Chung, Alexander Anderson. Evolutionary Tumor Board: Implementing dynamic personalized therapy using evolutionary theory and mathematical modeling for clinical decision support [abstract]. In: Proceedings of the AACR Special Conference on the Evolutionary Dynamics in Carcinogenesis and Response to Therapy; 2022 Mar 14-17. Philadelphia (PA): AACR; Cancer Res 2022;82(10 Suppl):Abstract nr PR010.

Abstract B025: Antifragile therapy

We present a novel paradigm of evolutionary cancer therapy based on the “antifragility” of the drug dose-response function. Antifragility is the word originally coined to describe the opposite of fragility. Systems or organisms can be described as antifragile if they derive a benefit from systemic variability, volatility, randomness, or disorder. In this work, we quantify the evolutionary and ecological benefit (or harm) derived from increasing dose volatility in treatment scheduling. Nonlinear sigmoidal dose-response curves are ubiquitous in medicine and have both convex and concave regions. Despite the ubiquity of dose response assays in biological sciences, these curves are typically used to measure differential response in first-order effects (mean value of drug dose delivered), while second-order effects (variance of drug dose) are generally ignored. Analysis of the convexity of dose response curves provides a direct prediction of response to continuous treatment (“even” schedules with zero volatility) in comparison to high-dose/low-dose treatment (“uneven” schedules with high volatility). For example, if the dose response function is antifragile (concave) near a dose of ‘x’, continuous administration of x may have a less efficacious response compared to a regimen that switches equally between 120% of x and 80% of x, even though both regimens use the same total drug. Mathematical analysis of dose response curves in vitro for a H3122 ALK-positive non-small cell lung cancer (NSCLC) cell line predicts that evolved-resistance cell lines can be more effectively treated using volatile treatment scheduling regimens, while treatment-naïve cell lines are most effectively treated by continuous treatment. However, selection pressure due to treatment selects for resistant phenotypes over time. We construct a mathematical model of gradual resistance, parameterized to data, and predict time-dependent antifragility in continuous (8 weeks), volatile (8 weeks) ALK inhibition in vivo. The key insight is that dose-response concavity (“anti-fragility”) increases in proportion to the amount of resistance in the tumor population. Antifragility provides a time-dependent metric which 1) predicts the emergence of resistance and 2) determines the optimal subsequent dosing strategy. Previous work indicates that resistance to ALK inhibitors occurs gradually, through the acquisition of multiple cooperating genetic and epigenetic adaptive changes. This observation led us to hypothesize that there is a critical point in the evolution of ALK-positive tumors where it is optimal to switch from continuous treatment to volatile dosing to optimally control the onset of gradual resistance. This hypothesis is also tested in vivo, comparing continuous and volatile treatment schedules of ALK inhibitors to a switching schedule of continuous-volatile (4 weeks each). We end by discussing the implications for adaptive therapy.

Citation Format: Jeffrey West, Bina Desai, Maximilian Strobl, Luke Pierik, Richard Miles, Cole Armagost, Mark Robertson-Tessi, Andriy Marusyk, Alexander R. A. Anderson. Antifragile therapy [abstract]. In: Proceedings of the AACR Special Conference on the Evolutionary Dynamics in Carcinogenesis and Response to Therapy; 2022 Mar 14-17. Philadelphia (PA): AACR; Cancer Res 2022;82(10 Suppl):Abstract nr B025.

Abstract B009: Integrating mathematical modelling and wet-lab experiments to examine the scope for adaptive treatment scheduling of PARP inhibitors in ovarian cancer

PARP inhibitors (PARPis) represent a great advancement in the treatment of ovarian cancer, yet these drugs often fail after a few months due to emerging drug resistance. A recent clinical trial in prostate cancer showed that evolutionary-inspired, adaptive drug scheduling significantly delayed time to progression. This approach adaptively skipped treatment to maintain a pool of drug-sensitive cells that suppressed resistant cells through competition. Here, we present results from a combined modelling and experimental study in which we investigated whether adaptive therapy could delay resistance to the PARPi olaparib in ovarian cancer.We performed a series of in vitro experiments in which we used Incucyte Zoom time-lapse microscopy to characterize the cell population dynamics under different PARPi schedules. Leveraging these data we developed an ordinary differential equation mathematical model of treatment response, and used this model to test different plausible adaptive treatment schedules. Our model can accurately predict the in vitro treatment dynamics, even to new schedules, and suggests that treatment modifications need to be carefully timed, or one risks losing control over tumor growth, even in the absence of any resistance. This is because multiple rounds of cell division are required for cells to acquire sufficient DNA damage to induce apoptosis. As a result, adaptive therapy algorithms that modulate treatment but never completely withdraw it are predicted to perform better in this setting than strategies based on treatment interruptions. Subsequent experiments confirm this prediction in vivo. To conclude, we present preliminary work aiming at investigating the scope for clinical translation of our results by confronting our model with longitudinal CA-125 data from 53 ovarian cancer patients receiving olaparib at Moffitt. Overall, this study contributes to a better understanding of the impact of scheduling on treatment outcome for PARPis, and showcases some of challenges involved in developing adaptive therapies for new treatment settings.

Citation Format: Maximilian Strobl, Mehdi Damaghi, Alexandra Martin, Samantha Byrne, Mark Robertson-Tessi, Robert Gatenby, Robert Wenham, Philip Maini, Alexander R.A. Anderson. Integrating mathematical modelling and wet-lab experiments to examine the scope for adaptive treatment scheduling of PARP inhibitors in ovarian cancer [abstract]. In: Proceedings of the AACR Special Conference on the Evolutionary Dynamics in Carcinogenesis and Response to Therapy; 2022 Mar 14-17. Philadelphia (PA): AACR; Cancer Res 2022;82(10 Suppl):Abstract nr B009.

Abstract B006: Evolutionary Tumor Board: Implementing dynamic personalized therapy using evolutionary theory and mathematical modeling for clinical decision support

This abstract is being presented as a short talk in the scientific program. A full abstract is available in the Proffered Abstracts section (PR010) of the Conference Proceedings.

Citation Format: Mark Robertson-Tessi, Joel Brown, Maria Poole, Kimberly Luddy, Andriy Marusyk, Jill Gallaher, Jeffrey West, Matthew Johnson, Heiko Enderling, Rikesh Makanji, Joaquim Farinhas, Robert Gatenby, Damon Reed, Christine Chung, Alexander Anderson. Evolutionary Tumor Board: Implementing dynamic personalized therapy using evolutionary theory and mathematical modeling for clinical decision support [abstract]. In: Proceedings of the AACR Special Conference on the Evolutionary Dynamics in Carcinogenesis and Response to Therapy; 2022 Mar 14-17. Philadelphia (PA): AACR; Cancer Res 2022;82(10 Suppl):Abstract nr B006.

Abstract PR008: Immunosuppressive niche engineering at the onset of human colorectal cancer

The evolutionary dynamics of tumor initiation remain undetermined, and the interplay between neoplastic cells and the immune system is hypothesized to be critical in transformation. Colorectal cancer (CRC) presents a unique opportunity to study the transition to malignancy as pre-cancers (adenomas) and early-stage cancers are frequently detected and surgically removed. Here, we demonstrate a key role for the immune response in tumor initiation by studying tumor-immune eco-evolutionary dynamics from pre-cancer to carcinoma using a computational model, ecological analysis of digital pathology data, and multi-region exome sequencing and neoantigen prediction in a total of 62 patient samples. Modelling indicates there are several potential routes to malignancy, each of which uniquely sculpts tumor ecology and intra-tumor antigenic heterogeneity (aITH). In patient samples, the immune microenvironment was characterized using the spatial distribution of 17 markers across registered whole-slide images, as well as patterns of intra-lesion aITH measured using multi-region exome sequencing and neoantigen prediction. The patient data were best described by a model where adenomas that become immunogenic early on do not progress to CRC because they are under immune control; progression therefore proceeds in adenomas with low immunogenicity. In these tumors, immune suppression is initially low, but gradually an immunosuppressive niche that is depleted in CD8+ cytotoxic T cells expands. There was little evidence for immune blockade (PD-L1 expression) in tumor initiation or progression. These results suggest that re-engineering the immunosuppressive niche may prove to be an effective immunotherapy in CRC.

Citation Format: Chandler D. Gatenbee, Ann-Marie Baker, Ryan O. Schenck, Maximilian Strobl, Jeffrey West, Mark Robertson-Tessi, Trevor A. Graham, Alexander R.A. Anderson. Immunosuppressive niche engineering at the onset of human colorectal cancer [abstract]. In: Proceedings of the AACR Special Conference on the Evolutionary Dynamics in Carcinogenesis and Response to Therapy; 2022 Mar 14-17. Philadelphia (PA): AACR; Cancer Res 2022;82(10 Suppl):Abstract nr PR008.

Abstract B008: Man vs machine: Crowdsourcing vs deep learning of adaptive therapy strategies

Cancer Crusade is a citizen science game in which players design treatment strategies to influence a virtual tumor growth model. The treatment regimens that the players construct are collected and sent to an online database for analysis, which may give insights into clinically effective adaptive strategies. In parallel, a machine learning-based approach was applied to the game to find optimal treatment policies, and the optimized therapy approaches generated are compared to those generated by the human players. For the core game engine in Cancer Crusade, we started with a previously published hybrid cellular automata model of tumor metabolism and growth. The original model was used to study how tumor cells evolve acid-mediated invasion and the impact of treatment on these metabolic processes. We expanded the model by adding more drugs and cell phenotypes, including drug-resistant cells. The game was released on mobile platforms and has been generating data from plays for several years. The treatment strategies were analyzed using dendrogram clustering to find key decision differences and profile their relative performance. We also performed network analysis to observe treatment transitions. These analyses indicated several effective adaptive strategies, which tended to oscillate between chemotherapy and a pro-angiogenic drug, or chemotherapy and a hypoxia-activated prodrug, or combinations of all three of these drugs. These results parallel a machine learning (Q-learning) approach to the problem, which yielded a preferred strategy based on chemotherapy combined with a pro-angiogenic drug. The existence of several human-discovered alternative strategies suggests that in general human players may offer greater variety of successful strategies via citizen science games than Q-learning. The human and machine learning approaches also excelled in different areas. Human players tended to use less drug overall, and perform best on less aggressive tumors, while the Q-learning approach used more drug overall and performed best on more aggressive tumors.

Citation Format: Rafael R. Bravo, Brenden K. Peterson, Mark Robertson-Tessi, Daniel M. Faissol, Alexander R.A. Anderson. Man vs machine: Crowdsourcing vs deep learning of adaptive therapy strategies [abstract]. In: Proceedings of the AACR Special Conference on the Evolutionary Dynamics in Carcinogenesis and Response to Therapy; 2022 Mar 14-17. Philadelphia (PA): AACR; Cancer Res 2022;82(10 Suppl):Abstract nr B008.

Abstract B023: An integrated approach to understanding the evolutionary dynamics of childhood acute lymphoblastic leukemia from diagnosis to relapse

Up to 20% of children with acute lymphoblastic leukemia (cALL) will relapse after initially responding to treatment. Dissecting the evolutionary population dynamics leading to relapse would help explain treatment failure from a mechanistic standpoint, aiding the design of more effective therapies. Comparisons of genetic heterogeneity at diagnosis and relapse have shown that relapse is often dominated by either a specific diagnostic subclone or its evolutionary progeny, leading to the idea that selection during treatment of cALL primarily operates at the genotype level. However, due to the technical difficulties associated with analyzing the rare cells that survive treatment in patients - definitive support for this idea is thus far missing. To overcome this challenge, we have previously developed a xenograft model of cALL induction chemotherapy treatment. Combining this with single-cell resolution analysis, we showed that, despite a massive reduction in leukemic burden, the first 28 days of chemotherapy have little impact on the genetic heterogeneity of cALL. This finding was inconsistent with the idea of selection acting at the level of genotypes. Instead, treatment induced a bottleneck at the level of cell state, determining the survival of a transcriptionally homogeneous population broadly characterized by reduced biosynthetic activity and cell dormancy. However, cALL treatment lasts several years and cannot be entirely modelled in vivo. Hence to assess whether genetic selection could act on a larger timescale or whether the clonal dominance frequently observed at relapse results from stochastic sweeps, we have implemented a data-driven mathematical model using the Hybrid Automata Library (HAL) to simulate longer treatment courses. The model allows for explicit spatial and temporal tracking of the evolutionary trajectories of individual cALL cells from diagnosis to relapse. Surprisingly we found that preserved genetic heterogeneity post-induction treatment and clonal dominance at relapse are features of virtually all relapsed leukemias; regardless of whether subclones with equal or varied fitness populate the diagnostic disease. This finding highlights the misinterpretation risks associated with limited disease snapshot analysis. Crucially, although genetically driven leukemias and leukemias in which all subclones have a similar probability of entering dormancy had similar endpoints, their temporal evolutionary dynamics largely differed. In the latter, reproducibly fewer cells survived induction chemotherapy, and relapse occurred on longer timelines, predominantly post-treatment. This observation provides the first empirical evidence of the notion that early and late relapse in cALLs may result from distinct selection mechanisms. Our preliminary data further suggest that even when high-fitness subclones are present, specifically targeting them is, in many cases, unlikely to improve overall outcome. Alternative dose fractionation protocols leveraging the epigenetically homogenous nature of residual cells may hold a better promise.

Citation Format: Virginia A. Turati, Javier Herrero Sanchez, Jeffrey West, Mark Robertson-Tessi, Tariq Enver, Andriy Marusyk, Alexander R. A. Anderson. An integrated approach to understanding the evolutionary dynamics of childhood acute lymphoblastic leukemia from diagnosis to relapse [abstract]. In: Proceedings of the AACR Special Conference on the Evolutionary Dynamics in Carcinogenesis and Response to Therapy; 2022 Mar 14-17. Philadelphia (PA): AACR; Cancer Res 2022;82(10 Suppl):Abstract nr B023.

Abstract B003: Evolutionary games in radiotherapy resistance: Exploiting an immunologic double bind in advanced prostate cancer

Evolution-informed therapy exploits the ecological and evolutionary consequences of drug resistance to inhibit the expansion of treatment resistant populations prolonging time to progression. One strategy, termed an evolutionary double bind, uses an initial therapy to elicit a specific adaptive mechanism that is then selectively targeted by a follow-on therapy. Typically, the second therapy is most effective in the recurrent setting and less effective when given upfront. Here we examine combining radiation therapy followed by immunotherapy as a double bind strategy. Radiotherapy (RT) can induce lethal double-strand DNA breaks in tumor cells and radiosensitivity is governed by the cells’ DNA repair capabilities. Radioresistant cancer cells upregulate DNA damage response pathways, which are mediated by ataxia telangiectasia mutated (ATM) and ataxia telangiectasia and Rad3-related (ATR) proteins. However, successful adaptations to RT and other DNA damaging agents can increase natural killer (NK) cell ligand expression on tumor cells increasing their vulnerability to the immune system. We evolved two cell lines, derived from a single population (22Rv1), with differing RT sensitivities (RT-sensitive and RT-resistant). We demonstrate RT-resistant cells upregulate NK cell ligands, including major histocompatibility complex class I chain-related protein A/B (MICA/B), poliovirus receptor (PVR), and PVRL2, resulting in a 2-fold increase in sensitivity to NK cell mediated killing. We conducted a multidisciplinary investigation of this potential evolutionary double bind through in vitro studies of radiation-sensitive and radiation-resistant cells treated with RT, NK cell therapy alone, or in combinations to parameterize evolution-based mathematical models. Mathematical modeling quantified three potential aspects of a double-bind: cost of resistance, inter-specific competition between RT-sensitive and RT-resistant lines, and preferential targeting of RT-resistant lines by NK cells. Despite a slower intrinsic growth rate, the RT-resistant population out competed the RT-sensitive cells in co-cultures at all seeding frequencies in the absence of treatment. Radiotherapy slowed overall growth but strongly selected for RT-resistant cells. NK cell therapy alone suppressed the RT-resistant population but maintained a residual population of radiation-sensitive cells. RT followed by NK cells was the most effective at reducing overall tumor burden, even in cases with an initially large RT-resistant fraction. We conclude RT followed by immunotherapy produces an evolutionary double bind that can be exploited in heterogenous tumors to limit RT resistance. Furthermore, model simulations predict extinction of both populations by sequential RT and NK cell therapy is achievable.

Citation Format: Kimberly A. Luddy, Jeffery West, Mark Robertson-Tessi, Alexander R.A. Anderson, Taylor M. Bursell, Laure Marignol, Cliona O'Farrelly, Robert Gatenby. Evolutionary games in radiotherapy resistance: Exploiting an immunologic double bind in advanced prostate cancer [abstract]. In: Proceedings of the AACR Special Conference on the Evolutionary Dynamics in Carcinogenesis and Response to Therapy; 2022 Mar 14-17. Philadelphia (PA): AACR; Cancer Res 2022;82(10 Suppl):Abstract nr B003.

Abstract B020: Distinct spatiotemporal tumor-immune ecologies enable disease prediction in NSCLC patients

Purpose of study: Non-small cell lung cancer (NSCLC) is the most common and fatal of cancers. In this work, we examine multiplexed images of NSCLC tumors to investigate both the spatial and spatiotemporal eco-evolutionary interactions between the tumor and its microenvironment, to better understand NSCLC tumor progression and therapy response. Methods: We developed a scalable, computational image analysis pipeline using cell-segmentation and quadrat-based approaches to analyze the spatial and temporal features of high-dimensional multiplexed NSCLC images. Multiplexed images enable the spatial readouts of numerous biomarkers per tissue sample and allow the interpretation of cellular states and the characterization of tumor-immune interactions across tissue ensembles. We also implement statistical approaches for ecological niche modelling combined with machine learning and deep learning models to predict disease progression and identify clinical imaging biomarkers. Data: Images were obtained from two 9-patient cohorts having advanced/metastatic NSCLC who were treated with the oral HDAC inhibitor vorinostat combined with the PD-1 inhibitor pembrolizumab. The first cohort had 4 progressors (PD) and 5 with stable disease (SD). The second cohort had 3 patients each in the PD, SD and partial response (PR) categories. Images were collected from all patients both pre- and on-treatment (during the third week). Results: Using our computational framework based on cell segments and quadrats, we confirm that different spatial neighborhoods exist that distinguish PD from SD, and that these spatial ecologies aid disease progression: PD patients have distinct ecologies with higher colocalization of PanCK+PD-1+FoxP3 indicating an immunosuppressive environment, whereas SD patients have a higher colocalization of PanCK+PD-L1 along with T cells, suggesting immunoactive tumor regions. These can be considered as potential biomarker candidates for predicting tumor progression. In an additional experiment where we include PR samples in our analyses, these distinct spatial neighborhoods are reinforced amongst PD, SD and PR patient groups corroborating the existence of spatiotemporal patterns. Further, we were able to predict treatment response with >91% accuracy given each patient’s spatial distribution of cell types from pre-treatment images. Using these predictions, we can generate risk maps at the patient level to identify areas of the tumor that are indicators of a higher probability of progression during treatment. Conclusions: We leveraged both single-cell and quadrat-resolution analysis of multiplexed imaging data and identified fundamentally distinct spatial ecologies between PD and SD patients. The ecology in PD patients appears to be primed for immune resistance even before treatment. This ecological diversity between SD and PD patients acts as a biomarker that enables accurate disease progression prediction. Our disease progression predictions can be used in conjunction with standard PD-L1 status to further strengthen personalized treatment strategies.

Citation Format: Sandhya Prabhakaran, Chandler Gatenbee, Mark Robertson-Tessi, Amer A. Beg, Jhanelle Gray, Scott Antonia, Robert A. Gatenby, Alexander R.A. Anderson. Distinct spatiotemporal tumor-immune ecologies enable disease prediction in NSCLC patients [abstract]. In: Proceedings of the AACR Special Conference on the Evolutionary Dynamics in Carcinogenesis and Response to Therapy; 2022 Mar 14-17. Philadelphia (PA): AACR; Cancer Res 2022;82(10 Suppl):Abstract nr B020.

Abstract B033: Immunosuppressive niche engineering at the onset of human colorectal cancer

This abstract is being presented as a short talk in the scientific program. A full abstract is available in the Proffered Abstracts section (PR0081) of the Conference Proceedings.

Citation Format: Chandler D. Gatenbee, Ann-Marie Baker, Ryan O. Schenck, Maximilian Strobl, Jeffrey West, Mark Robertson-Tessi, Trevor A. Graham, Alexander R.A. Anderson. Immunosuppressive niche engineering at the onset of human colorectal cancer [abstract]. In: Proceedings of the AACR Special Conference on the Evolutionary Dynamics in Carcinogenesis and Response to Therapy; 2022 Mar 14-17. Philadelphia (PA): AACR; Cancer Res 2022;82(10 Suppl):Abstract nr B033.

Fluctuating methylation clocks for cell lineage tracing at high temporal resolution in human tissues

Molecular clocks that record cell ancestry mutate too slowly to measure the short-timescale dynamics of cell renewal in adult tissues. Here, we show that fluctuating DNA methylation marks can be used as clocks in cells where ongoing methylation and demethylation cause repeated 'flip-flops' between methylated and unmethylated states. We identify endogenous fluctuating CpG (fCpG) sites using standard methylation arrays and develop a mathematical model to quantitatively measure human adult stem cell dynamics from these data. Small intestinal crypts were inferred to contain slightly more stem cells than the colon, with slower stem cell replacement in the small intestine. Germline APC mutation increased the number of replacements per crypt. In blood, we measured rapid expansion of acute leukemia and slower growth of chronic disease. Thus, the patterns of human somatic cell birth and death are measurable with fluctuating methylation clocks (FMCs).

Immunosuppressive niche engineering at the onset of human colorectal cancer

The evolutionary dynamics of tumor initiation remain undetermined, and the interplay between neoplastic cells and the immune system is hypothesized to be critical in transformation. Colorectal cancer (CRC) presents a unique opportunity to study the transition to malignancy as pre-cancers (adenomas) and early-stage cancers are frequently resected. Here, we examine tumor-immune eco-evolutionary dynamics from pre-cancer to carcinoma using a computational model, ecological analysis of digital pathology data, and neoantigen prediction in 62 patient samples. Modeling predicted recruitment of immunosuppressive cells would be the most common driver of transformation. As predicted, ecological analysis reveals that progressed adenomas co-localized with immunosuppressive cells and cytokines, while benign adenomas co-localized with a mixed immune response. Carcinomas converge to a common immune "cold" ecology, relaxing selection against immunogenicity and high neoantigen burdens, with little evidence for PD-L1 overexpression driving tumor initiation. These findings suggest re-engineering the immunosuppressive niche may prove an effective immunotherapy in CRC.

Roadmap on plasticity and epigenetics in cancer

The role of plasticity and epigenetics in shaping cancer evolution and response to therapy has taken center stage with recent technological advances including single cell sequencing. This roadmap article is focused on state-of-the-art mathematical and experimental approaches to interrogate plasticity in cancer, and addresses the following themes and questions: is there a formal overarching framework that encompasses both non-genetic plasticity and mutation-driven somatic evolution? How do we measure and model the role of the microenvironment in influencing/controlling non-genetic plasticity? How can we experimentally study non-genetic plasticity? Which mathematical techniques are required or best suited? What are the clinical and practical applications and implications of these concepts?

The role of memory in non-genetic inheritance and its impact on cancer treatment resistance

Intra-tumour heterogeneity is a leading cause of treatment failure and disease progression in cancer. While genetic mutations have long been accepted as a primary mechanism of generating this heterogeneity, the role of phenotypic plasticity is becoming increasingly apparent as a driver of intra-tumour heterogeneity. Consequently, understanding the role of this plasticity in treatment resistance and failure is a key component of improving cancer therapy. We develop a mathematical model of stochastic phenotype switching that tracks the evolution of drug-sensitive and drug-tolerant subpopulations to clarify the role of phenotype switching on population growth rates and tumour persistence. By including cytotoxic therapy in the model, we show that, depending on the strategy of the drug-tolerant subpopulation, stochastic phenotype switching can lead to either transient or permanent drug resistance. We study the role of phenotypic heterogeneity in a drug-resistant, genetically homogeneous population of non-small cell lung cancer cells to derive a rational treatment schedule that drives population extinction and avoids competitive release of the drug-tolerant sub-population. This model-informed therapeutic schedule results in increased treatment efficacy when compared against periodic therapy, and, most importantly, sustained tumour decay without the development of resistance.

Abstract PO-044: Exploiting a radiotherapy induced immunologic double bind in advanced prostate cancer

An evolutionary double bind occurs in cancer when resistance to a commonly used therapy increases the sensitivity to a second therapy. Typically, the second therapy is most effective in the recurrent setting and less effective when given up front. There is a tremendous need for novel second-line therapies able to exploit these new sensitivities. Prostate cancer remains one of the most treatable cancers, however, advanced disease is typically metastatic and currently incurable. Radiation therapy (RT) is often used in the resistant setting. Given locally, RT induces double-strand DNA breaks in tumor cells. This results in cellular stress responses that either trigger apoptosis and cell death, or DNA repair and cell survival. Radiosensitivity is related to the DNA repair capabilities of a tumor. Activation of DNA damage response pathways, which are mediated by ataxia telangiectasia mutated (ATM) and ataxia telangiectasia and Rad3-related protein (ATR), induces changes in surviving cancer cells. Most notably, it has been shown that RT, as well as other DNA damaging agents, can increase natural killer cell (NK) ligand expression on tumor cells. Here we utilize an isogenic cell line model of radioresistant prostate cancer. Our pre-clinical model shows that alterations in DNA repair pathways required to maintain resistance to RT correlate with changes in NK-cell ligands, including: major histocompatibility complex class I chain-related protein A/B (MICA/B), Nectin-2 (CD112), and poliovirus receptor (PVR). This results in a 2-fold increase in sensitivity to NK-cell mediated killing in the radiation resistant cell line. An “evolutionary game assay” was performed by preparing mixtures of parental and radiation-resistant cell lines at 4 different ratios: 100:0, 90:10, 10:90 and 0:100. This was repeated for six treatment scenarios: untreated, 6Gy radiation, 1:1 NK cells, 5:1 NK cells, 6Gy with 1:1 NK cells, and 6Gy with 5:1 NK cells. Comparison of monotypic growth rates to coculture mixtures of parental and radiation-resistant cell lines enable quantification of competitive interactions through the use of evolutionary math models. This assay quantifies effective fitness (growth rate) to experimentally determine the frequency-dependent gradient of selection. Preliminary results indicate the emergence of three qualitatively unique evolutionary scenarios: 1) dominance of resistance (radiation therapy), 2) dominance of sensitive (5:1 NK), and 3) coexistence (untreated; 1:1 NK). These results suggest the existence of an evolutionary double-bind when NK-cell administration follows radiation. This was experimentally validated and confirmed by repeating the evolutionary assay with sequential radiation and NK therapy. Importantly, there is a minimum threshold of NK cells required to achieve the double-bind. Direct measurement of the fitness landscape underlying radiation and NK treatment enabled the discovery and validation of an evolutionary strategy to re-sensitize tumors after radiation treatment.

Citation Format: Kimberly Luddy, Jeffrey West, Mark Robertson-Tessi, Alexander Anderson, Laure Marignol, Robert Gatenby, Cliona O'Farrelly. Exploiting a radiotherapy induced immunologic double bind in advanced prostate cancer [abstract]. In: Proceedings of the AACR Virtual Special Conference on Radiation Science and Medicine; 2021 Mar 2-3. Philadelphia (PA): AACR; Clin Cancer Res 2021;27(8_Suppl):Abstract nr PO-044.

Normal tissue architecture determines the evolutionary course of cancer

Cancer growth can be described as a caricature of the renewal process of the tissue of origin, where the tissue architecture has a strong influence on the evolutionary dynamics within the tumor. Using a classic, well-studied model of tumor evolution (a passenger-driver mutation model) we systematically alter spatial constraints and cell mixing rates to show how tissue structure influences functional (driver) mutations and genetic heterogeneity over time. This approach explores a key mechanism behind both inter-patient and intratumoral tumor heterogeneity: competition for space. Time-varying competition leads to an emergent transition from Darwinian premalignant growth to subsequent invasive neutral tumor growth. Initial spatial constraints determine the emergent mode of evolution (Darwinian to neutral) without a change in cell-specific mutation rate or fitness effects. Driver acquisition during the Darwinian precancerous stage may be modulated en route to neutral evolution by the combination of two factors: spatial constraints and limited cellular mixing. These two factors occur naturally in ductal carcinomas, where the branching topology of the ductal network dictates spatial constraints and mixing rates.

Turnover Modulates the Need for a Cost of Resistance in Adaptive Therapy

Adaptive therapy seeks to exploit intratumoral competition to avoid, or at least delay, the emergence of therapy resistance in cancer. Motivated by promising results in prostate cancer, there is growing interest in extending this approach to other neoplasms. As such, it is urgent to understand the characteristics of a cancer that determine whether or not it will respond well to adaptive therapy. A plausible candidate for such a selection criterion is the fitness cost of resistance. In this article, we study a general, but simple, mathematical model to investigate whether the presence of a cost is necessary for adaptive therapy to extend the time to progression beyond that of a standard-of-care continuous therapy. Tumor cells were divided into sensitive and resistant populations and we model their competition using a system of two ordinary differential equations based on the Lotka-Volterra model. For tumors close to their environmental carrying capacity, a cost was not required. However, for tumors growing far from carrying capacity, a cost may be required to see meaningful gains. Notably, it is important to consider cell turnover in the tumor, and we discuss its role in modulating the impact of a resistance cost. To conclude, we present evidence for the predicted cost-turnover interplay in data from 67 patients with prostate cancer undergoing intermittent androgen deprivation therapy. Our work helps to clarify under which circumstances adaptive therapy may be beneficial and suggests that turnover may play an unexpectedly important role in the decision-making process. SIGNIFICANCE: Tumor cell turnover modulates the speed of selection against drug resistance by amplifying the effects of competition and resistance costs; as such, turnover is an important factor in resistance management via adaptive therapy.See related commentary by Strobl et al., p. 811.

Searching for Goldilocks: How Evolution and Ecology Can Help Uncover More Effective Patient-Specific Chemotherapies

Deaths from cancer are mostly due to metastatic disease that becomes resistant to therapy. A mainstay treatment for many cancers is chemotherapy, for which the dosing strategy is primarily limited by patient toxicity. While this MTD approach builds upon the intuitively appealing principle that maximum therapeutic benefit is achieved by killing the largest possible number of cancer cells, there is increasing evidence that moderation might allow host-specific features to contribute to success. We believe that a "Goldilocks Window" of submaximal chemotherapy will yield improved overall outcomes. This window combines the complex interplay of cancer cell death, immune activity, emergence of chemoresistance, and metastatic dissemination. These multiple activities driven by chemotherapy have tradeoffs that depend on the specific agents used as well as their dosing levels and schedule. Here we present evidence supporting the idea that MTD may not always be the best approach and offer suggestions toward a more personalized treatment regime that integrates insights into patient-specific eco-evolutionary dynamics.

Evolutionary dynamics of neoantigens in growing tumors

Cancers accumulate mutations that lead to neoantigens, novel peptides that elicit an immune response, and consequently undergo evolutionary selection. Here we establish how negative selection shapes the clonality of neoantigens in a growing cancer by constructing a mathematical model of neoantigen evolution. The model predicts that, without immune escape, tumor neoantigens are either clonal or at low frequency; hypermutated tumors can only establish after the evolution of immune escape. Moreover, the site frequency spectrum of somatic variants under negative selection appears more neutral as the strength of negative selection increases, which is consistent with classical neutral theory. These predictions are corroborated by the analysis of neoantigen frequencies and immune escape in exome and RNA sequencing data from 879 colon, stomach and endometrial cancers.

The Immune Checkpoint Kick Start: Optimization of Neoadjuvant Combination Therapy Using Game Theory

PURPOSE

In an upcoming clinical trial at the Moffitt Cancer Center for women with stage 2/3 estrogen receptor-positive breast cancer, treatment with an aromatase inhibitor and a PD-L1 checkpoint inhibitor combination will be investigated to lower a preoperative endocrine prognostic index (PEPI) that correlates with relapse-free survival. PEPI is fundamentally a static index, measured at the end of neoadjuvant therapy before surgery. We have developed a mathematical model of the essential components of the PEPI score to identify successful combination therapy regimens that minimize tumor burden and metastatic potential, on the basis of time-dependent trade-offs in the system.

METHODS

We considered two molecular traits, CCR7 and PD-L1, which correlate with treatment response and increased metastatic risk. We used a matrix game model with the four phenotypic strategies to examine the frequency-dependent interactions of cancer cells. This game was embedded in an ecological model of tumor population-growth dynamics. The resulting model predicts evolutionary and ecological dynamics that track with changes in the PEPI score.

RESULTS

We considered various treatment regimens on the basis of combinations of the two therapies with drug holidays. By considering the trade off between tumor burden and metastatic potential, the optimal therapy plan was a 1-month kick start of the immune checkpoint inhibitor followed by 5 months of continuous combination therapy. Relative to a protocol giving both therapeutics together from the start, this delayed regimen resulted in transient suboptimal tumor regression while maintaining a phenotypic constitution that is more amenable to fast tumor regression for the final 5 months of therapy.

CONCLUSION

The mathematical model provides a useful abstraction of clinical intuition, enabling hypothesis generation and testing of clinical assumptions.

Hybrid Automata Library: A flexible platform for hybrid modeling with real-time visualization

The Hybrid Automata Library (HAL) is a Java Library developed for use in mathematical oncology modeling. It is made of simple, efficient, generic components that can be used to model complex spatial systems. HAL's components can broadly be classified into: on- and off-lattice agent containers, finite difference diffusion fields, a GUI building system, and additional tools and utilities for computation and data collection. These components are designed to operate independently and are standardized to make them easy to interface with one another. As a demonstration of how modeling can be simplified using our approach, we have included a complete example of a hybrid model (a spatial model with interacting agent-based and PDE components). HAL is a useful asset for researchers who wish to build performant 1D, 2D and 3D hybrid models in Java, while not starting entirely from scratch. It is available on GitHub at https://github.com/MathOnco/HAL under the MIT License. HAL requires the Java JDK version 1.8 or later to compile and run the source code.

Model genotype-phenotype mappings and the algorithmic structure of evolution

Cancers are complex dynamic systems that undergo evolution and selection. Personalized medicine approaches in the clinic increasingly rely on predictions of tumour response to one or more therapies; these predictions are complicated by the inevitable evolution of the tumour. Despite enormous amounts of data on the mutational status of cancers and numerous therapies developed in recent decades to target these mutations, many of these treatments fail after a time due to the development of resistance in the tumour. The emergence of these resistant phenotypes is not easily predicted from genomic data, since the relationship between genotypes and phenotypes, termed the genotype-phenotype (GP) mapping, is neither injective nor functional. We present a review of models of this mapping within a generalized evolutionary framework that takes into account the relation between genotype, phenotype, environment and fitness. Different modelling approaches are described and compared, and many evolutionary results are shown to be conserved across studies despite using different underlying model systems. In addition, several areas for future work that remain understudied are identified, including plasticity and bet-hedging. The GP-mapping provides a pathway for understanding the potential routes of evolution taken by cancers, which will be necessary knowledge for improving personalized therapies.

Acidity promotes tumour progression by altering macrophage phenotype in prostate cancer

BACKGROUND

Tumours rapidly ferment glucose to lactic acid even in the presence of oxygen, and coupling high glycolysis with poor perfusion leads to extracellular acidification. We hypothesise that acidity, independent from lactate, can augment the pro-tumour phenotype of macrophages.

METHODS

We analysed publicly available data of human prostate cancer for linear correlation between macrophage markers and glycolysis genes. We used zwitterionic buffers to adjust the pH in series of in vitro experiments. We then utilised subcutaneous and transgenic tumour models developed in C57BL/6 mice as well as computer simulations to correlate tumour progression with macrophage infiltration and to delineate role of acidity.

RESULTS

Activating macrophages at pH 6.8 in vitro enhanced an IL-4-driven phenotype as measured by gene expression, cytokine profiling, and functional assays. These results were recapitulated in vivo wherein neutralising intratumoural acidity reduced the pro-tumour phenotype of macrophages, while also decreasing tumour incidence and invasion in the TRAMP model of prostate cancer. These results were recapitulated using an in silico mathematical model that simulate macrophage responses to environmental signals. By turning off acid-induced cellular responses, our in silico mathematical modelling shows that acid-resistant macrophages can limit tumour progression.

CONCLUSIONS

This study suggests that tumour acidity contributes to prostate carcinogenesis by altering the state of macrophage activation.

The Goldilocks Window of Personalized Chemotherapy: Getting the Immune Response Just Right

The immune system is a robust and often untapped accomplice of many standard cancer therapies. A majority of tumors exist in a state of immune tolerance where the patient's immune system has become insensitive to the cancer cells. Because of its lymphodepleting effects, chemotherapy has the potential to break this tolerance. To investigate this, we created a mathematical modeling framework of tumor-immune dynamics. Our results suggest that optimal chemotherapy scheduling must balance two opposing objectives: maximizing tumor reduction while preserving patient immune function. Successful treatment requires therapy to operate in a "Goldilocks Window" where patient immune health is not overly compromised. By keeping therapy "just right," we show that the synergistic effects of immune activation and chemotherapy can maximize tumor reduction and control. SIGNIFICANCE: To maximize the synergy between chemotherapy and antitumor immune response, lymphodepleting therapy must be balanced in a "Goldilocks Window" of optimal dosing.Graphical Abstract: http://cancerres.aacrjournals.org/content/canres/79/20/5302/F1.large.jpg.

Abstract B17: A dual in vivo and in silico system to model ectopic lymph node structure formation and antitumor immune response in the murine tumor microenvironment

The tumor microenvironment is a dynamic and complex system consisting of the tumor bed, surrounding stroma, vasculature, and immune infiltrate that self-organizes in response to contextual queues and polarizing factors. The nature and extent of this organization can determine if an immune response is beneficial, suppressed, or tumor promoting. Previous efforts have identified a gene expression signature (GES) that is composed of twelve chemokines and that is associated with the presence of ectopic lymph node structures (ELNS) within the tumor microenvironment. These ELNS are predictive of increased survival across a spectrum of solid tumors, and are thought to be foci of active antigen presentation, lymphocyte recruitment, and immune activation. Understanding how and why these structures form is paramount for improving existing immunotherapies and identifying novel treatment modalities. Most in vivo models utilize subcutaneous orthotopic inoculations that, while producing a tumorous mass, lack the surrounding context of a physiologically relevant microenvironment and completely lack ELNS. Alternatively, naturally occurring or chemically induced models offer a more physiologically relevant setting, but are often highly variable systems in which controlling experimental parameters becomes problematic. Here, we build upon insights gained from the GES using an implantable three-dimensional bioscaffold system to study the interaction of a developing tumor with endogenous immune infiltrates and implanted stroma. This allows for the introduction of carefully controlled microenvironmental elements in a more consistent mouse model. In this system, we can selectively interrogate individual chemokines using slow-release microparticles, and visualize immune infiltrates in real time using intravital microscopy. Concurrently, we use this system to parameterize an integrated mathematical model of the tumor microenvironment, which can then reinform our three-dimensional bioscaffold model. This dual-model system has identified an important role for three components of the GES--CCL19, CCL21, and CXCL13--in the induction of ELNS, and suggests that the type and extent of stromal activation can greatly enhance or prevent effective antitumor immunity. In particular, stromal activation using the lymphotoxin beta receptor ligand, LTα1β2, can induce organized aggregates of endogenous lymphocytes; significantly increase infiltration of T cells, B cells, and dendritic cells (all p<0.05); and prevent the growth of MC-38 colon carcinoma cells (p<0.001 by implant weight). In silico model runs parameterized by ex vivo microtaxis assays and in vivo ELNS formation assays predict that polarized activation of stromal cells is sufficient to compartmentalize lymphoid aggregates into discrete B cell and T cell zones and promote antitumor activity. Conversely, stromal activation in the presence of certain type-II promoting factors, such as all-trans retinoic acid, can result in the infiltration of suppressive populations that promote tumorigenesis. Taken together, this novel dual-model system of the tumor microenvironment suggests that the context of stromal activation in the tumor microenvironment can either promote effective antitumor immunity by inducing ELNS, or suppress antitumor immune response by recruiting suppressive populations.

Citation Format: Adam W. Mailloux, Falahat Rana, Yohsuke Yagawa, Mark Robertson-Tessi, Zhou L. Susan, Alexander R. A. Anderson, James J. Mulé. A dual in vivo and in silico system to model ectopic lymph node structure formation and antitumor immune response in the murine tumor microenvironment [abstract]. In: Proceedings of the AACR Special Conference: Advances in Modeling Cancer in Mice: Technology, Biology, and Beyond; 2017 Sep 24-27; Orlando, Florida. Philadelphia (PA): AACR; Cancer Res 2018;78(10 Suppl):Abstract nr B17.

Defining Cancer Subpopulations by Adaptive Strategies Rather Than Molecular Properties Provides Novel Insights into Intratumoral Evolution

Ongoing intratumoral evolution is apparent in molecular variations among cancer cells from different regions of the same tumor, but genetic data alone provide little insight into environmental selection forces and cellular phenotypic adaptations that govern the underlying Darwinian dynamics. In three spontaneous murine cancers (prostate cancers in TRAMP and PTEN mice, pancreatic cancer in KPC mice), we identified two subpopulations with distinct niche construction adaptive strategies that remained stable in culture: (i) invasive cells that produce an acidic environment via upregulated aerobic glycolysis; and (ii) noninvasive cells that were angiogenic and metabolically near-normal. Darwinian interactions of these subpopulations were investigated in TRAMP prostate cancers. Computer simulations demonstrated invasive, acid-producing (C2) cells maintain a fitness advantage over noninvasive, angiogenic (C3) cells by promoting invasion and reducing efficacy of immune response. Immunohistochemical analysis of untreated tumors confirmed that C2 cells were invariably more abundant than C3 cells. However, the C2 adaptive strategy phenotype incurred a significant cost due to inefficient energy production (i.e., aerobic glycolysis) and depletion of resources for adaptations to an acidic environment. Mathematical model simulations predicted that small perturbations of the microenvironmental extracellular pH (pHe) could invert the cost/benefit ratio of the C2 strategy and select for C3 cells. In vivo, 200 mmol/L NaHCO₃ added to the drinking water of 4-week-old TRAMP mice increased the intraprostatic pHe by 0.2 units and promoted proliferation of noninvasive C3 cells, which remained confined within the ducts so that primary cancer did not develop. A 0.2 pHe increase in established tumors increased the fraction of C3 cells and signficantly diminished growth of primary and metastatic tumors. In an experimental tumor construct, MCF7 and MDA-MB-231 breast cancer cells were coinjected into the mammary fat pad of SCID mice. C2-like MDA-MB-231 cells dominated in untreated animals, but C3-like MCF7 cells were selected and tumor growth slowed when intratumoral pHe was increased. Overall, our data support the use of mathematical modeling of intratumoral Darwinian interactions of environmental selection forces and cancer cell adaptive strategies. These models allow the tumor to be steered into a less invasive pathway through the application of small but selective biological force. Cancer Res; 77(9); 2242-54. ©2017 AACR.

Abscopal Benefits of Localized Radiotherapy Depend on Activated T-cell Trafficking and Distribution between Metastatic Lesions

It remains unclear how localized radiotherapy for cancer metastases can occasionally elicit a systemic antitumor effect, known as the abscopal effect, but historically, it has been speculated to reflect the generation of a host immunotherapeutic response. The ability to purposefully and reliably induce abscopal effects in metastatic tumors could meet many unmet clinical needs. Here, we describe a mathematical model that incorporates physiologic information about T-cell trafficking to estimate the distribution of focal therapy-activated T cells between metastatic lesions. We integrated a dynamic model of tumor-immune interactions with systemic T-cell trafficking patterns to simulate the development of metastases. In virtual case studies, we found that the dissemination of activated T cells among multiple metastatic sites is complex and not intuitively predictable. Furthermore, we show that not all metastatic sites participate in systemic immune surveillance equally, and therefore the success in triggering the abscopal effect depends, at least in part, on which metastatic site is selected for localized therapy. Moreover, simulations revealed that seeding new metastatic sites may accelerate the growth of the primary tumor, because T-cell responses are partially diverted to the developing metastases, but the removal of the primary tumor can also favor the rapid growth of preexisting metastatic lesions. Collectively, our work provides the framework to prospectively identify anatomically defined focal therapy targets that are most likely to trigger an immune-mediated abscopal response and therefore may inform personalized treatment strategies in patients with metastatic disease.

Abstract A19: Systems biology approach predicts the diagnostic value of T effector: T regulatory cell ratio in clinical response to combined radiation/immunotherapy of high-risk soft tissue sarcoma

Background: The combination of radiotherapy and immunotherapy for cancer treatment has regained momentum since the recent remarkable clinical success in immune checkpoint strategies. The attraction comes from the concept that radiation-induced tumor cell death will release tumor antigens capable of activating dendritic cells to establish long term adaptive T cell immunity against the tumor. However, such combination trials are difficult to interpret, especially when proper markers are not available that predict clinical response. Using a set of data published on patients with high risk soft tissue sarcoma who were treated with intratumoral administration of autologous dendritic cells and local fractionated external beam radiation, we developed a mathematical model examining the levels of T effector cells and T regulatory cells that were reported in each patient who did or did not demonstrate a clinical response.

Materials and Methods: The Phase II clinical trial included 15 Grade 2 high-risk soft tissue sarcoma patients who received radiation/dendritic cell therapy, of which there were 5 responders and 10 non-responders. With the T cell phenotypes that were found in the patients and the accompanying clinical outcome, we designed a mathematical model of tolerogenic and immunogenic tumor subpopulations and their interactions with the host immune system, comprised of T effector and T regulatory cells. Tumor response to the investigational radio-immunotherapy protocol is simulated. The model is calibrated to fit patient-specific pre- and post-treatment response dynamics using computational genetic algorithms. Cell kinetics that separate responder and non-responder cohorts are also categorized.

Results: The mathematical model can be calibrated to reproduce patient-specific tumor volume and immune cell number evolution during treatment. Comparison of responder and non-responder cohorts reveals that immune T effector cell recruitment and efficacy are determinants of treatment response. In contrast, tumor growth dynamics were indistinguishable between individual patients and patient cohorts. Increased T effector to T regulatory cell ratio at diagnosis (2.74 in responders vs. 1.88 in non-responders) emerges as a prognostic marker for treatment response.

Conclusion: A calibrated quantitative tumor model may help to identify mechanisms that determine treatment response. Current clinical practice predicts treatment outcome predominantly based on tumor characteristics. Our initial work indicates that tumor growth dynamics are indistinguishable between high-risk soft tissue sarcoma patients that respond to radio-immunotherapy and those who do not. Instead, our preliminary studies suggest that the host response to the growing tumor – the quantity and quality of the immune response in particular – may serve exclusively as a prognostic marker. This suggests a departure from the current paradigm to derive prognostic factors from genetic characterization of tumor biopsy samples.

Citation Format: Sotiris Prokopiou, Jan Poleszczuk, Mark Robertson-Tessi, Kimberly A. Luddy, Mayer Fishman, Eduardo Moros, Julie Y. Djeu, Heiko Enderling. Systems biology approach predicts the diagnostic value of T effector: T regulatory cell ratio in clinical response to combined radiation/immunotherapy of high-risk soft tissue sarcoma. [abstract]. In: Proceedings of the AACR Special Conference: Tumor Immunology and Immunotherapy: A New Chapter; December 1-4, 2014; Orlando, FL. Philadelphia (PA): AACR; Cancer Immunol Res 2015;3(10 Suppl):Abstract nr A19.

Abstract A86: Harnessing T-cell homeostasis to diagnose and treat solid and liquid tumors

T-cell populations are subject to homeostatic control from cytokines and microenvironmental signaling. Disruption of homeostasis can cause changes to the dynamics of the system that have implications for the progression of cancer. Here we present two mathematical models that examine the progression of tumors in the context of T-cell homeostasis and provide therapeutic decision support in the clinic. Model 1: During a chronic disease such as cancer, T cells often become tolerant to the antigens presented by the disease. This tolerant state effectively limits the response of the immune system to the tumor. Experimental evidence has shown that depletion of T-cells can lead to a loss of T-cell tolerance. During the homeostatic phase of T-cell compartment repopulation, there is a temporary window of opportunity during which T cells lose their tolerant state, allowing them to respond to tumor antigens. In addition, clonal expansion of the tumor-specific T-cell clone may be enhanced during the regrowth phase due to increased stimulation. We use an ordinary differential equation (ODE) model to explore the effect of T-cell depletion and homeostatic repopulation on the loss of tolerance in the T-cell compartment and subsequent effectiveness of immune-mediated tumor cytotoxicity. The model predicts different outcomes for the tumor and T-cell compartment, dependent on the strength and schedule of the depletion therapy. The optimal regimen can lead to tumor control in some cases, but T-cell exhaustion is also common dynamic predicted by the model. By understanding the effects of T-cell depletion, immune depleting therapies can be optimized to enhance immune potential. Model 2: Large Granular Lymphocytic Leukemia (LGLL) is a T-cell lymphoproliferative disorder that exhibits clonal expansion of a subset of T cells. Since there are no clinical biomarkers to predict the aggressiveness of the disease, treatment decisions are often made on a watch and wait approach. Using a set of ODEs, we develop a model of LGLL that uses clinical patient data from diagnosis to predict the timeframe for progression of the disease. Our experimental results have suggested that the disease is caused by a change in sensitivity to both positive and negative regulators of T-cell homeostasis. The model incorporates these cell-specific mechanisms to investigate their effect when placed in a homeostatic setting. The level of dysregulation as measured from patient-specific data determines the rate of outgrowth of the diseased T-cell clone, and therefore serve as a useful predictive tool for managing treatment decisions in the clinic.

Citation Format: Mark Robertson-Tessi, Derek Park, Kimberly Luddy, Adam Mailloux, Pearlie K. Burnette, Alexander Anderson. Harnessing T-cell homeostasis to diagnose and treat solid and liquid tumors. [abstract]. In: Proceedings of the AACR Special Conference: Tumor Immunology and Immunotherapy: A New Chapter; December 1-4, 2014; Orlando, FL. Philadelphia (PA): AACR; Cancer Immunol Res 2015;3(10 Suppl):Abstract nr A86.

Impact of metabolic heterogeneity on tumor growth, invasion, and treatment outcomes

Histopathologic knowledge that extensive heterogeneity exists between and within tumors has been confirmed and deepened recently by molecular studies. However, the impact of tumor heterogeneity on prognosis and treatment remains as poorly understood as ever. Using a hybrid multiscale mathematical model of tumor growth in vascularized tissue, we investigated the selection pressures exerted by spatial and temporal variations in tumor microenvironment and the resulting phenotypic adaptations. A key component of this model is normal and tumor metabolism and its interaction with microenvironmental factors. The metabolic phenotype of tumor cells is plastic, and microenvironmental selection leads to increased tumor glycolysis and decreased pH. Once this phenotype emerges, the tumor dramatically changes its behavior due to acid-mediated invasion, an effect that depends on both variations in the tumor cell phenotypes and their spatial distribution within the tumor. In early stages of growth, tumors are stratified, with the most aggressive cells developing within the interior of the tumor. These cells then grow to the edge of the tumor and invade into the normal tissue using acidosis. Simulations suggest that diffusible cytotoxic treatments, such as chemotherapy, may increase the metabolic aggressiveness of a tumor due to drug-mediated selection. Chemotherapy removes the metabolic stratification of the tumor and allows more aggressive cells to grow toward blood vessels and normal tissue. Antiangiogenic therapy also selects for aggressive phenotypes due to degradation of the tumor microenvironment, ultimately resulting in a more invasive tumor. In contrast, pH buffer therapy slows down the development of aggressive tumors, but only if administered when the tumor is still stratified. Overall, findings from this model highlight the risks of cytotoxic and antiangiogenic treatments in the context of tumor heterogeneity resulting from a selection for more aggressive behaviors.

A model for effects of adaptive immunity on tumor response to chemotherapy and chemoimmunotherapy

Complete clinical regressions of solid tumors in response to chemotherapy are difficult to explain by direct cytotoxicity alone, because of low growth fractions and obstacles to drug delivery. A plausible indirect mechanism that might reconcile this is the action of the immune system. A model for interaction between tumors and the adaptive immune system is presented here, and used to examine controllability of tumors through the interplay of cytotoxic, cytostatic and immunogenic effects of chemotherapy and the adaptive immune response. The model includes cytotoxic and helper T cells, T regulatory cells (Tregs), dendritic cells, memory cells, and several key cytokines. Nearly all parameter estimates are derived from experimental and clinical data. Individual tumors are characterized by two parameters: growth rate and antigenicity, and regions of tumor control are identified in this parameter space. The model predicts that inclusion of the immune response significantly expands the region of tumor control for both cytostatic and cytotoxic chemotherapies. Moreover, outside the control zone, tumor growth is delayed significantly. An optimal fractionation schedule is predicted, for a fixed cumulative dose. The model further predicts expanded regions of tumor control when several forms of immunotherapy (adoptive T cell transfer, Treg depletion, and dendritic cell vaccination) are combined with chemotherapy. Outcomes depend greatly on tumor characteristics, the schedule of administration, and the type of immunotherapy chosen, suggesting promising opportunities for personalized medicine. Overall, the model provides insight into the role of the adaptive immune system in chemotherapy, and how scheduling and immunotherapeutic interventions might improve efficacy.

Abstract B05: Harnessing heterogeneity to design better combination therapies

Background: Significant cellular, molecular, and tissue heterogeneity is widely observed between and within tumors and the potential clinical significance of these variations is increasingly recognized. The intricate dialogue between the tumor cells and their environment selects for tumor phenotypes that are best adapted to survive. However, this environment is temporally and spatially heterogeneous largely due to variations in blood flow which results in local fluctuations of nutrients, growth factors and other cellular populations. This variability can lead to a heterogeneous response of the tumor to a given therapy. Interactions that occur within the cancer ecosystem do so in a dynamic spatio-temporal manner that is almost impossible to dissect via experimentation alone, so we propose a theoretical framework to generate combination therapies that are sensitive to initial tumor heterogeneity.

Methods: Using a hybrid multi-scale mathematical model of tumor growth in vascularized tissue, we investigate the selection pressures exerted by spatial and temporal variations in tumor microenvironment and the resulting phenotypic adaptations. A key component of this model is normal and tumor metabolism and its interaction with microenvironmental factors. The metabolic phenotype of tumor cells is plastic, and microenvironmental selection leads to increased tumor glycolysis and decreased pH. Once this phenotype emerges, the tumor dramatically changes its behavior due to acid-mediated invasion, an effect that depends on the heterogeneity of the tumor cell phenotypes and their spatial distribution within the tumor. The tumors grown within this in silico model display much phenotypic variation, and this heterogeneity depends on the conditions of the microenvironment and the plasticity of the tumor cells.

Results: Using the model, we generate sets of tumors with different biological parameters, classify them according to their spatial and temporal heterogeneity, and then administer several therapies, including chemotherapy, vascular therapy, pH buffer therapy, and hypoxia-activated drugs. The model predicts that pH buffer therapy will only have a tumor-preventative effect if administered before the tumor acquires the heterogeneous state that leads to acid-mediated invasion. This is in agreement with experimental results from a spontaneous prostate tumor mouse model (TRAMP mouse). In general, the model predicts that the outcomes of each therapy are highly dependent on the initial tumor heterogeneity at the time of commencing treatment. We categorize the ‘signatures’ of each therapy outcome as a function of the heterogeneity class of the initial tumor. By understanding the signature of each drug in isolation, we implement drug combinations in a sequence that promotes synergistic response for a given class of tumor heterogeneity. The signature of the first drug in the sequence is used to pick the following complementary drug. This produces a more intelligent treatment regimen that can be designed to harness tumor heterogeneity and modulate its impact on treatment outcomes.

Citation Format: Mark Robertson-Tessi, Robert J. Gillies, Robert A. Gatenby, Alexander RA Anderson. Harnessing heterogeneity to design better combination therapies. [abstract]. In: Abstracts: AACR Special Conference on Cellular Heterogeneity in the Tumor Microenvironment; 2014 Feb 26-Mar 1; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2015;75(1 Suppl):Abstract nr B05. doi:10.1158/1538-7445.CHTME14-B05

The role of toll-like receptors in colorectal cancer progression: evidence for epithelial to leucocytic transition

Toll-like receptors (TLRs) are expressed by immune cells, intestinal epithelium, and tumor cells. In the homeostatic setting, they help to regulate control over invading pathogens and maintain the epithelial lining of the large and small intestines. Aberrant expression of certain TLRs by tumor cells can induce growth inhibition while others contribute to tumorigenesis and progression. Activation of these TLRs can induce inflammation, tumor cell proliferation, immune evasion, local invasion, and distant metastasis. These TLR-influenced behaviors have similarities with properties observed in leukocytes, suggesting that tumors may be hijacking immune programs to become more aggressive. The concept of epithelial to leucocytic-transition (ELT) is proposed, akin to epithelial to mesenchymal transition, in which tumors develop the ability to activate leucocytic traits otherwise inaccessible to epithelial cells. Understanding the mechanisms of ELT could lead to novel therapeutic strategies for inhibiting tumor metastasis.

Abstract A2: The importance of metabolic heterogeneity and its consequences on tumor invasion, metastatic growth, and treatment

Heterogeneity of both the tumor microenvironment and cellular phenotype plays a key role in guiding tumor behavior. Metabolism acts as a central integrator to mediate cellular response to the changing environment, linking the effects of hypoxia, acidosis, proliferation and necrosis. Using a hybrid multi-scale mathematical model of tumor growth in a vascularized tissue, we investigate the selection pressures exerted by the tumor microenvironment as the cancer progresses. A key focus is the metabolism of both normal and tumor cells, and how small changes in phenotype can lead to the development of more metabolically aggressive groups of cells. Development of a glycolytic phenotype due to microenvironmental selection pressures leads to acid-mediated invasion. Results from the model suggest that these aggressive cells have a dramatic advantage as a metastatic seed, compared to less aggressive seeds from the same primary. In addition, there are significant differences in metastatic growth rate depending on the soil, suggesting that prevention or treatment of metastases in different organs may require different strategies. Application of pH buffering therapy has been shown in mouse models to slow down or prevent both primary tumor and metastatic growth. This treatment has been simulated in the model, and the results show that the effect of the treatment is highly sensitive to the timing of application. Additional results suggest that diffusible cytotoxic treatments such as chemotherapy may increase the metabolic aggressiveness of a tumor post-treatment due to the altered selection pressure caused by the drug. Chemotherapy removes the metabolic stratification of the tumor and allows more aggressive cells to grow towards blood vessels and normal tissue. A third treatment we consider is the use of anti-angiogenic therapy, which foments the development of aggressive phenotypes due to degradation of the tumor microenvironment. The model highlights the role of metabolic heterogeneity and microenvironmental selection in driving tumor invasion and metastatic development.

Citation Format: Mark Robertson-Tessi, Robert J. Gillies, Robert A. Gatenby, Alexander R.A. Anderson. The importance of metabolic heterogeneity and its consequences on tumor invasion, metastatic growth, and treatment. [abstract]. In: Proceedings of the AACR Special Conference on Tumor Invasion and Metastasis; Jan 20-23, 2013; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2013;73(3 Suppl):Abstract nr A2.

Abstract 53: Exploiting heterogeneity to develop better treatment strategies using an evolutionary multiscale mathematical model of tumor-vessel interactions

Tumor heterogeneity in cancer is an observed fact, both genetically and phenotypically. Cell-cell variation is seen in almost all aspects of cancer from early development all the way through to invasion and subsequent metastasis. The tumor microenvironment is also very heterogeneous and dynamic with multiple zones of hypoxia, normoxia, acidosis and necrosis. Central to our understanding of this heterogeneity is how the tumor cells interact with each other and with their microenvironment. Since the microenvironment is modulated by the tumor itself, as well as the surrounding stroma and the vasculature supply, these interactions rapidly become complex. In order to better understand these complex interactions, we have developed an integrated approach utilizing a hybrid multiscale mathematical model in parallel with an experimental murine window chamber model. This study aims to understand the development of the vasculature in response to a growing tumor, the effects of a heterogeneous microenvironment, and how interactions between the tumor and its vasculature modulate tumor progression and treatment resistance. By incorporating phenotypic heterogeneity into the model, a range of tumor phenotypes emerges due to the heterogeneous microenvironment. The response of the tumor to a treatment will depend on the distribution of phenotypes as well as the local conditions. At the same time, treatment will alter the phenotypic distribution and the microenvironment, so that subsequent tumor progression will be profoundly different than it was prior to treatment. By using the spatial and temporal projections of the mathematical model, optimal treatment strategies are derived and classified based on tumor and microenvironmental parameters. One particularly intriguing result is that tumor control, as opposed to eradication, may be more readily achieved by modulation of the tumor heterogeneity via treatment-mediated microenvironmental selection.

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 53. doi:10.1158/1538-7445.AM2011-53