Multiple model-informed open-loop control of uncertain intracellular signaling dynamics.

Computational approaches to tune the activation of intracellular signal transduction pathways both predictably and selectively will enable researchers to explore and interrogate cell biology with unprecedented precision. Techniques to control complex nonlinear systems typically involve the application of control theory to a descriptive mathematical model. For cellular processes, however, measurement assays tend to be too time consuming for real-time feedback control and models offer rough approximations of the biological reality, thus limiting their utility when considered in isolation. We overcome these problems by combining nonlinear model predictive control with a novel adaptive weighting algorithm that blends predictions from multiple models to derive a compromise open-loop control sequence. The proposed strategy uses weight maps to inform the controller of the tendency for models to differ in their ability to accurately reproduce the system dynamics under different experimental perturbations (i.e. control inputs). These maps, which characterize the changing model likelihoods over the admissible control input space, are constructed using preexisting experimental data and used to produce a model-based open-loop control framework. In effect, the proposed method designs a sequence of control inputs that force the signaling dynamics along a predefined temporal response without measurement feedback while mitigating the effects of model uncertainty. We demonstrate this technique on the well-known Erk/MAPK signaling pathway in T cells. In silico assessment demonstrates that this approach successfully reduces target tracking error by 52% or better when compared with single model-based controllers and non-adaptive multiple model-based controllers. In vitro implementation of the proposed approach in Jurkat cells confirms a 63% reduction in tracking error when compared with the best of the single-model controllers. This study provides an experimentally-corroborated control methodology that utilizes the knowledge encoded within multiple mathematical models of intracellular signaling to design control inputs that effectively direct cell behavior in open-loop.

Bone marrow-derived progenitor cells augment venous remodeling in a mouse dorsal skinfold chamber model.

The delivery of bone marrow-derived cells (BMDCs) has been widely used to stimulate angiogenesis and arteriogenesis. We identified a progenitor-enriched subpopulation of BMDCs that is able to augment venular remodeling, a generally unexplored area in microvascular research. Two populations of BMDCs, whole bone marrow (WBM) and Lin(-)/Sca-1(+) progenitor cells, were encapsulated in sodium alginate and delivered to a mouse dorsal skinfold chamber model. Upon observation that encapsulated Sca-1(+) progenitor cells enhance venular remodeling, the cells and tissue were analyzed on structural and molecular levels. Venule walls were thickened and contained more nuclei after Sca-1(+) progenitor cell delivery. In addition, progenitors expressed mRNA transcript levels of chemokine (C-X-C motif) ligand 2 (CXCL2) and interferon gamma (IFNγ) that are over 5-fold higher compared to WBM. Tissues that received progenitors expressed significantly higher protein levels of vascular endothelial growth factor (VEGF), monocyte chemotactic protein-1 (MCP-1), and platelet derived growth factor-BB (PDGF-BB) compared to tissues that received an alginate control construct. Nine days following cell delivery, tissue from progenitor recipients contained 39% more CD45(+) leukocytes, suggesting that these cells may enhance venular remodeling through the modulation of the local immune environment. Results show that different BMDC populations elicit different microvascular responses. In this model, Sca-1(+) progenitor cell-derived CXCL2 and IFNγ may mediate venule enlargement via modulation of the local inflammatory environment.