Biological machines / microfludics

Recent work has demonstrated the pervasive role of mechanics in biology. The most prominent examples include matrix stiffness influencing stem cell lineage and tumor progression, axonal tension regulating presynaptic vesicle clustering, and stiffness gradients guiding migrating cells. Mechanotransduction, the mechanism underlying these cellular responses to mechanical stimuli, has been studied in detail for the past decade and offers a new paradigm for directing the form and function of integrated cellular systems.

Over the past 5 years, we developed various microfluidic platforms for mimicking the three dimensional microenvironment and investigating the role of mechanical stimuli, such as interstitial flow, cyclic strain, and ECM stiffness gradients, on cellular processes including cell migration, angiogenesis, and differentiation.

Recently, we have drawn upon our understanding of mechanobiology to direct the function of multicellular systems. For example, we extended our angiogenesis model to build functional vascular networks in vitro, and we directed stem cell differentiation into cardiomyocytes by applying cyclic strain. As we increase the complexity of synthetic modules toward building biological machines, mechanics will play a more significant role, particularly in the engineering of neurons and myocytes for sensing and actuation. We will employ mechanical engineering as a tool to address this complexity while simultaneously extending our understanding of mechanotransduction.
Vivek Sivathanu
Jordy Whisler.

Angiogenesis/ Vasculogenesis

  Formation of new blood vessel from an existing branch, by a regulated process known as angiogenesis, governs vascular patterning in the body and determines the distribution of nutrients and oxygen supply. Angiogenesis has essential roles in development, reproduction and repair but also occurs in tumor formation and in a variety of diseases [1, 2]. Our lab studies the angiogenic process by computational modeling across multiple scales [3, 4] and by in vitro microfluidic experiments that mimics in vivo biophysical and biochemical microenvironment. We showed that angiogenic endothelial cells seeded in contact with collagen gel can be induced to form nascent angiogenic sprouts in microfluidic which later develop into a vascular network [5, 6, 7].

To understand the single cell decisions in angiogenesis at the signaling level, we model individual cell as a decision making entities and follow individual cell as they make decisions in angiogenic conditions [8]. In collaboration with the Lauffenburger lab at MIT, we attempt to elucidate how such single cell decision might be governed by an intracellular signaling by measuring the intracellular changes in signaling activities upon stimulating cells with potent factors that induce and suppress sprout formation [9].
Investigators: Michelle Chen, Jordy Whisler, Ran Li, Vivek Sivathanu, Anya Burkart


Tumor invasion has received a lot of attention as a critical step in metastatic disease for developing new cancer drugs. Current understanding of the role of biophysical and cellular microenvironment in tumor invasion is limited, because of the lack of appropriate in vitro and in vivo models. We have adapted our previous microfluidic platforms [1] for studying the role of the endothelium on tumor intravasation and the effects of interstitial flow on tumor cell migration, along with the development of new hard plastic devices for commercial transition. Recent results from the tumor-endothelial interaction assay demonstrated the capability to form a 3D endothelium on collagen type I matrices, in the presence of invading tumor cells in 3D (Figure 1). Upon stimulation with inflammatory cytokines we demonstrated an increase in diffusive permeability to fluorescent dextrans, in agreement with a measured increase in the number of intravasation events. These results demonstrate the utility of this assay for studying the role of the endothelial barrier function in tumor cell intravasation.
We developed a microfluidic system for investigating the role of interstitial flow in tumor cell migration (Figure 2). Tumor cells exposed to interstitial flow preferentially migrated along streamlines, and the relative percentage of cells migrating upstream and downstream is a function of chemokine receptor activity and cell density. Interstitial flow stimulates downstream tumor cell migration through CCR7 autocrine signaling. However, flow also stimulates upstream cell migration through a competing, mechanically mediated pathway, as evidenced by significantly increased FAK activation in devices with flow. The relative strength of the autocrine and mechanical stimuli determines whether cells migrate upstream or downstream of the flow direction.


We applied the known commercially-viable manufacturing methods to a cyclic olefin copolymer (COC) material to fabricate a microfluidic device with controlled surface properties and improved potential to serve high-volume applications. Culture of cells in the new hard plastic device indicated no adverse effects of the COC material. Therefore, this transition of platform demonstrates a capability of using microfludic devices for 3D cell culture across the range from the scientific research to applications with broad clinical impact.
Investigators: Ran Li, Michelle Chen, Alexandra Boussommier-Calleja, Anya Burkart.


Simulation and modeling

Schematic of the angiogenesis process within a microfluidic device. A stochastic cell population model was used with experimental data gathered from the microfluidic platform to develop a closed-loop control system model for angiogenesis (From ref. 4).
Computational models aide with data interpretation and experimental design, and simulations can prove insight into biological mechanisms in instances where experiments are not feasible. Modeling and simulation are integral parts to the Mechanobiology Lab, and we have developed models spanning length scales from single molecules to cell populations. Furthermore, these models are not independent; we employ course-graining techniques to allow models developed at small length scales to inform larger scale models. For example, the bulk properties of a material have been estimated by course-graining simulations of the constitutive atoms, providing a quantitative link between the chemical composition and mechanics of biomaterials (1). The actin cytoskeleton contributes to the mechanical rigidity of cells, and dynamic reorganization of the intracellular actin network is required for key cellular events such as proliferation and migration. Mechanical force plays a crucial role in governing the dynamics of the actin cytoskeleton, and we have developed a computational model derived from Brownian dynamics simulations to study the mechanical properties of actin networks (2). The model well captures experimentally observed viscoelastic properties of actin and provides novel insight into the contribution of molecular motors and actin crosslinking proteins to the viscoelastic moduli of actin networks. Recently, we have extended this model to investigate the role of molecular motors in the generation of actin stress fibers, molecular complexes under tension that provide direct mechanical connections to the extracellular environment.    
  In the MIT Mechanobiology Lab, our experiments are tightly coupled with computational models for investigating biological phenomena. The highly controlled cell microenvironment enabled by our microfluidic platforms allows validation of our cell-level models, which in turn, provide insight into the mechanism underlying experimentally observed cellular behavior. We have integrated the microfluidic platform for studying angiogenesis with a hybrid discrete-continuum model to investigate the effects of matrix and growth factors on vascular network topology (3), and we have explored the link between matrix degrading enzymes and angiogenic sprout structure through control theory (4). Recently we have developed finite element models for investigating the role of interstitial flow in perturbing the tumor cell microenvironment (5), and we are adapting these models to aide in the development of patterned synthetic tissues and biological machines.

Shear deformation of a computationally generated actin network. In silico actin networks were implemented to investigate the role of actin crosslinking proteins and molecular motors on the viscoelastic properties of physiologically relevant actin networks (From ref. 2).
Investigators: Michael Mak, Andrea Malandrino.

Simulation results from a hybrid continuum-discrete model show similar topology to neovascular networks generated in microfluidic devices (From ref. 3).