Parametric Model of Combination Therapy for Non-Hodgkin Lymphoma

Abstract

The development and clinical testing of drug combinations for the treatment of Non-Hodgkin Lymphoma (NHL) and other cancers has recently shown great promise. However, determining the optimum combination and its associated dosages for maximum efficacy and minimum side effects is still a challenge. This paper describes a parametric analysis of the dynamics of malignant B-cells and the effects of an anti-sense oligonucleotide targeted to BCL-2 (as-bcl-2), anti-CD-20 (rituximab) and their combination, for a SCID mouse human lymphoma xenograft model of NHL. Our parametric model is straightforward. Several mechanisms of malignant B-cell birth and death in the nodal micro-environment are simulated. Cell death is accelerated by hypoxia and starvation induced by tumor scale, by modification of anti-apoptosis with as-bcl-2, and by direct kill effects of rituximab (cell kill by cytotoxic immune cells is not included, due to the absence of an immune system in the corresponding experiments). We show that the cell population dynamics in the control animals are primarily determined by K*, the ratio of rate constants for malignant cell death, Kd, and cell birth, Kb. Tumor growth with independent treatments is reproduced by the model, and is used to predict their effect when administered in combination. Malignant cell lifetimes are derived to provide a quantitative comparison of the efficacy of these treatments. Future experimental and clinical applications of the model are discussed.

Immune Response to Combination Therapy for Non-Hodgkin Lymphomas

“…The previously developed model [1] equated the temporal rate of change of the malignant B-cell population (N B ) to the sum of terms that characterize the various mechanisms that increase or decrease the population in the absence of an immune response. These mechanisms include the normal B-cell birth rate, the malignant B-cell death rate, and the potential amplification of the malignant population death rate by hypoxia and lack of nutrients in the micro-environment (modeled to first-order by the ratio of the malignant B-cell population to its initial value).…”

“…The value of parameter K″′ appears to be less important, although it determines the rate of tumor regression. Figure 3 provides an illustrative example of the predicted behavior (compared to the SCID model [1] ) for a particular set of parameter values. B-cell regression is predicted to start at about 20 days, and the T-cell population becomes dominant at later times.…”

“…In a recent study [1] , we developed a parametric model of tumor growth in B-cell lymphomas to predict the quantitative effects of combination drug therapy on malignant cell population dynamics. Our model extends earlier theoretical studies of immune response to tumor growth [2] , [3] by deriving values of key rate constants from mono-therapy experiments that are then utilized to make combination therapy predictions.…”

“…We then apply this model to the cases previously studied, primarily as an illustration of the influence of an immune response with combination immunotherapy. A very brief review of the SCID data analysis previously reported [1] is provided, as we are focusing on the predicted effects of combination immunotherapy in an immune-competent mouse, rather than further statistical analysis of the published SCID data. Finally, we illustrate the application of our expanded model to optimization of drug combination dosages, drug scheduling, and the issue of indolent to aggressive disease transformation.…”

Monitoring Disease Progression and Therapeutic Response in a Disseminated Tumor Model for Non-Hodgkin Lymphoma by Bioluminescence Imaging

“…In the past, subcutaneous tumor models were established. [12][13][14] However, these models do not mirror human lymphoma adequately because lymphomas are typically disseminated in humans. Therefore, a disseminated xenograft tumor model would be more appropriate to study chemotherapy effects in a preclinical model.…”