Traulsen, Arne

Abstract Abstract 1236 Background: Although most patients with early chronic phase CML (ECP-CML) respond to TKI therapy, the depth and speed of response can be different. While it is known that both the Sokal and Hasford scores have an impact on the speed and depth of response, no mechanisms explaining these differences have been identified. Objective: To provide explanations for any potential differences in CML response as a function of the Sokal and Hasford score using a computational model. Methods: We utilize a computational model of hematopoiesis and CML, together with serial quantitative data of disease burden under nilotinib therapy to determine the fraction of CML cells responding to therapy (z) and the impact of TKI on the self-renewal probability of CML cells (e) under therapy. Patients were stratified at diagnosis on both the Sokal and Hasford scoring system. A non-linear least squares method was used to separately fit the model to serial Q-RT-PCR data for BCR-ABL in response to therapy in each cohort. Results: A total of 73 patients were studied. The number of patients with low, intermediate and high risk disease based on the Sokal score was 34, 29 and 10 respectively while the respective distribution of patients on the Hasford score was 29, 43 and 1. Although the impact of nilotinib on the self-renewal probability of CML progenitor cells was similar across all risk groups (there were substantial differences in the fraction of cells responding to therapy: For the Sokal groups, the fraction of cells (z) responding to therapy decreased from 0.09 to 0.086 and 0.069 respectively for low, intermediate and high risk disease. In the case of the Hasford score, the difference in z between low and intermediate categories becomes more pronounced, i.e. z is 0.093 for low and 0.08 for intermediate risk disease. Conclusions: The risk score at diagnosis of CML has a direct impact on the dynamics of response to TKI therapy. Patients with a lower risk score respond faster to the same therapy when compared to high risk patients. The impact of TKI on the self renewal of CML cells appears to be the same regardless of the risk score but the fraction of cells responding decreases as the risk score increases. This suggests that subclones that may be less sensitive to TKI therapy may be emerging. Strategies that increase the fraction of cells responding to therapy in patients with higher risk disease may be indicated. Disclosures: Rosti: Novartis: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Speakers Bureau; Bristol Myers Squibb: Honoraria, Speakers Bureau; Roche: Speakers Bureau.

The cellular age distribution of hierarchically organized tissues can reveal important insights into the dynamics of cell differentiation and self-renewal and associated cancer risks. Here, we examine the effect of progenitor compartments with varying differentiation and self-renewal capacities on the resulting observable distributions of replicative cellular ages. We find that strongly amplifying progenitor compartments, i.e. compartments with high self-renewal capacities, substantially broaden the age distributions which become skewed towards younger cells with a long tail of few old cells. For several of these strongly amplifying compartments, the age distribution becomes virtually independent of the influx from the stem cell compartment. By contrast, if tissues are organized into many downstream compartments with low self-renewal capacity, the shape of the replicative cell distribution in more differentiated compartments is dominated by stem cell dynamics with little added variation. In the limiting case of a strict binary differentiation tree without self-renewal, the shape of the output distribution becomes indistinguishable from that of the input distribution. Our results suggest that a comparison of cellular age distributions between healthy and cancerous tissues may inform about dynamical changes within the hierarchical tissue structure, i.e. an acquired increased self-renewal capacity in certain tumours. Furthermore, we compare our theoretical results to telomere length distributions in granulocyte populations of 10 healthy individuals across different ages, highlighting that our theoretical expectations agree with experimental observations.

AbstractCancer, the unregulated proliferation of cells, can occur at any age and may arise from almost all cell types. However, the incidence and types of cancer differ with age. Some cancers are predominantly observed in children, others are mostly restricted to older ages. Treatment strategies of some cancers are very successful and cure is common in childhood, while treatment of the same cancer type is much more challenging in adults. Here, we develop a stochastic model of stem cell proliferation that considers both tissue development and homeostasis and discuss the disturbance of such a system by mutations. Due to changes in population size, mutant fitness becomes context dependent and consequently the effects of mutations on the stem cell population can vary with age. We discuss different mutant phenotypes and show the age dependency of their expected abundances. Most importantly, fitness of particular mutations can change with age and advantageous mutations can become deleterious or vice versa. This perspective can explain unique properties of childhood disorders, for example, the frequently observed phenomenon of a self-limiting leukemia in newborns with trisomy 21, but also explains other puzzling observations such as the increased risk of leukemia in patients with bone marrow failure or chemotherapy induced myelodysplasia.

We investigate the in vivo patterns of stem cell divisions in the human hematopoietic system throughout life. In particular, we analyze the shape of telomere length distributions underlying stem cell behavior within individuals. Our mathematical model shows that these distributions contain a fingerprint of the progressive telomere loss and the fraction of symmetric cell proliferations. Our predictions are tested against measured telomere length distributions in humans across all ages, collected from lymphocyte and granulocyte sorted telomere length data of 356 healthy individuals, including 47 cord blood and 28 bone marrow samples. We find an increasing stem cell pool during childhood and adolescence and an approximately maintained stem cell population in adults. Furthermore, our method is able to detect individual differences from a single tissue sample, i.e. a single snapshot. Prospectively, this allows us to compare cell proliferation between individuals and identify abnormal stem cell dynamics, which affects the risk of stem cell related diseases.

Abstract Many tumors are believed to be maintained by a small number of cancer stem–like cells, where cure is thought to require eradication of this cell population. In this study, we investigated the dynamics of acute promyelocytic leukemia (APL) before and during therapy with regard to disease initiation, progression, and therapeutic response. This investigation used a mathematical model of hematopoiesis and a dataset derived from the North American Intergroup Study INT0129. The known phenotypic constraints of APL could be explained by a combination of differentiation blockade of PML–RARα–positive cells and suppression of normal hematopoiesis. All-trans retinoic acid (ATRA) neutralizes the differentiation block and decreases the proliferation rate of leukemic stem cells in vivo. Prolonged ATRA treatment after chemotherapy can cure patients with APL by eliminating the stem-like cell population over the course of approximately one year. To our knowledge, this study offers the first estimate of the average duration of therapy that is required to eliminate stem-like cancer cells from a human tumor, with the potential for the refinement of treatment strategies to better manage human malignancy. Cancer Res; 74(19); 5386–96. ©2014 AACR.

Cancers are rarely caused by single mutations, but often develop as a result of the combined effects of multiple mutations. For most cells, the number of possible cell divisions is limited because of various biological constraints, such as progressive telomere shortening, cell senescence cascades or a hierarchically organized tissue structure. Thus, the risk of accumulating cells carrying multiple mutations is low. Nonetheless, many diseases are based on the accumulation of such multiple mutations. We model a general, hierarchically organized tissue by a multi-compartment approach, allowing any number of mutations within a cell. We derive closed solutions for the deterministic clonal dynamics and the reproductive capacity of single clones. Our results hold for the average dynamics in a hierarchical tissue characterized by an arbitrary combination of proliferation parameters. We show that hierarchically organized tissues strongly suppress cells carrying multiple mutations and derive closed solutions for the expected size and diversity of clonal populations founded by a single mutant within the hierarchy. We discuss the example of childhood acute lymphoblastic leukaemia in detail and find good agreement between our predicted results and recently observed clonal diversities in patients. This result can contribute to the explanation of very diverse mutation profiles observed by whole genome sequencing of many different cancers.

Haemopoiesis is a fundamental physiologic process found in many animals. Among mammals, the diversity in size and function required suitable adaptations of this process. In this work, we use allometric principles to determine whether this required a change in the basic architecture of haemopoiesis. We show that it is possible to express both the number and rate with which haemopoietic stem cells replicate as well as total marrow output across all mammals as a function of adult mass. This unified view, which is compatible with the existing data, suggests that there was no need for major adaptations in the architecture of haemopoiesis across mammals.

Abstract The pool of hematopoietic stem cells that actively contributes to hematopoiesis is small, and the cells replicate slowly. Patients with paroxysmal nocturnal hemoglobinuria invariably have a mutation in the PIG-A gene, and many have more than one clone of PIG-A mutated cells. Typically there is a dominant clone and a smaller second clone. By using a combination of stochastic dynamics and models of hematopoiesis, we show that it is very unlikely that more than one PIG-A mutated clone arises at the level of the hematopoietic stem cells. More likely, the smaller clone develops in the progenitor cell pool that would be expected to contribute to hematopoiesis for a shorter period of time. We provide estimates for the duration of these contributions and testable hypotheses that can shed important insights on this acquired hematopoietic stem cell disorder. Disclosure of potential conflicts of interest is found at the end of this article.