Dynamics of Complex Biological Systems

At the convergence of physics with biology, our group is motivated both by the desire to gain fundamental insights into the behavior of living systems and by the drive to contribute to the pressing challenges associated with the explosion of quantitative information in medical research.

Our analysis of shape dynamics of migrating cells has led us to discover mechanical waves as a ubiquitous underlying motors in many fast-migrating cells. Another project in the Biodynamics lab is to elucidate how surface chemistry and topography affects migratory machinery, and how internal waves may be harnessed to control cell behavior. We also developed new tools to control the arrangement and dynamics of cell groups via holographic laser tweezers and to investigate the mechanical properties of models of circulating tumor cells. Other projects apply Complex Systems approaches to investigate cancer related biological processes as part of a Cancer Technology interaction between the University of Maryland and the National Cancer Institute. Examples of the many ways studying complex systems provides insight into biological systems are available on this page.

Actin Dynamics in Neuronal Model Systems, Neuronal Development, and Neuronal Network Activity

Kate O'Neill, Post-Doctoral Research Associate
with Dr. Edward Giniger (National Institutes of Health)

The dynamic nature of the cytoskeleton is essential to many cellular processes. As one of the most ubiquitous cytoskeletal proteins, actin is crucial to the development and maintenance of cell morphology, particularly in neurons, where morphological features often determine functional outcomes. The goal of this work is, therefore, to study the relationship between actin dynamics and electrical and biochemical signaling in systems of increasing complexity: “neuron-like” cells (human embryonic kidney cells [HEKs]), cultures of in vitro rat neurons, and in vivo wing discs of fruit fly pupae (Drosophila melanogaster). Using spinning disk confocal microscopy and high-resolution optical techniques this research will determine how actin dynamics contribute to the morphological development of neurons as well as the propagation of neuronal network activity at various time and length scales.

Neuronal Actin
Actin dynamics in the growth cone of a pioneer axon in the developing wing disc of Drosophila melanogaster.

Collective Migration during Cancer Progression

Rachel Lee, Postdoctoral Research Associate
with Dr. Stuart Martin (University of Maryland School of Medicine)
Rachel is supported by the T32 Cancer Biology Training Grant at the University of Maryland, Baltimore

In addition to playing a role in processes such as wound healing and development, collective migration is seen in the progression of diseases such as cancer. As tumor cells become more malignant, they gain the ability to migrate throughout the body; in addition to migrating as individual cells, they have been seen to migrate collectively in vivo and there is increasing evidence that collective behavior plays a role metastsis.

Using automated image analysis techniques we are able to extract information such as velocities from images of migrating human epithelial cells. Inspired by tools developed to study fluid flows and moving grains of sand, we have quantified the motion of migrating epithelial sheets and measured differences in malignant and non-malignant cells. We are currently using these tools to understand how the motion of epithelial cells is regulated and how changes in cell migration are linked to metastatic cancer.

Publications:
R.M. Lee, D.H. Kelley, K.N. Nordstrom, N.T. Ouellette, and W. Losert, New J Phys 2013 [IOP]
R.M. Lee, C.H. Stuelten, C.A. Parent, W. Losert, CSPO 2016 [IOP]
R.M. Lee, H. Yue, W.J. Rappel, W. Losert, Interface 2017 [The Royal Society]

Cell Sheet and PIV
Two images of a migrating sheet of MCF10A cells (top) compared to the velocity information derived using particle image velocimetry on these images (bottom).

Biocompatible Nano-topographical Surfaces

Matthew J. Hourwitz, PhD Student, Chemistry
with John Fourkas (University of Maryland)

Micro and nano-topographical surfaces are of great interest in our lab to probe mechanical perturbations to cell migration and actin dynamics. To fabricate these surfaces, master patterns are designed using multiphoton absorption polymerization (MAP). They can then be replicated using soft lithography and replica molding. We are working on ways to increase the scale of patterns and of output. The polymer replicas are biocompatible: we have observed overall cell survival on these surfaces for up to two months. The research also involves understanding how material and surface chemistry affect the cellular systems we probe.

Surface Fabrication Method
Design, molding, and large-scale production of micro/nanoscale topographical patterns.

Unidirectional Cell Guidance Promoted by Asymmetric Nanotopography

Xiaoyu Sun, PhD Alumna
with John Fourkas (University of Maryland)

Many biological and physiological processes depend upon directed migration of cells, which is typically mediated by chemical gradients, physical gradients, and signal relay. My work focuses on the effect that local topographic asymmetries on the nano/micro-scale have on attempting to guide Dictyostelium discoideum and lamellipod-driven human neutrophils to migrate in single, preferred directions. Given that these asymmetries can be repeated we can thereby attempt to provide directional guidance over arbitrarily large areas. We are currently studying the relationship between the direction and strength of the guidance and the details of the nano/microtopography. We have thus far demonstrated that asymmetric nano/microtopography guides the direction of internal actin polymerization waves and that cells move in the same direction as these waves. The conservation of this mechanism across cell types and the asymmetric shape of many natural scaffolds suggests that actin-wave-based guidance is important in biology and physiology.

Publication: X. Sun, et al., PNAS 2015 [PNAS]

Supported by the National Institutes of Health

Left: Sixty-frame (2.55-min) space/time plot of actin waves along a ridge composed of asymmetric sawteeth. Top, right: Actin-wave directionality angle histogram. Bottom, right: Average actin flux around a sawtooth.

Electrotaxis of Neutrophil-like Cells

Abby Bull, PhD Student, Physics
with Dr. Min Zhao (UC Davis) and Dr. Quan Qing (ASU)

Directed migration of cells are facilitated by topographic, chemotactic and electrostatic cues in order for biological functions to be carried out. Our lab has previously focused on the first two cues but now we are combining them with electrical gradients. Electric fields can attract neutrophils to facilitate in the healing process. They have also been seen to affect the migration of many other cell types including Dictyostelium discoideum and various types of tumorigenic cells. We hope to characterize the migration and actin dynamics of neutrophil-like HL 60 cells in response to a variety of electrical stimuli to better understand the mechanism of electrotaxis. The work focuses on the cues of both local topography and electric fields in complementing and competing set up with varied field strengths and length scales.

Migrating HL60 cell
Migration of a neutrophil-like HL60 cell.

Induced Differential Gene Expression via Nanotopographical Surfaces

Sylvester Gates, Post-Baccalaureate Faculty Assistant
with John Fourkas (University of Maryland)

We are working to understand the differentially expressed genes in cells undergoing esotaxis on ridged nano-topographical surfaces. Previous studies have shown electrotaxis and chemotaxis can induce changes in gene expression that differ from those in wild type or undisturbed conditions. To determine what common pathways underlie all these forms of cell migration and which pathways are unique to each form of guided migration, we are using gene expression of the transcriptome using Illumina next-generation sequencing (NGS). Further work is focused on testing actin imaging probes to be used in conjunction with voltage sensors to understand the effects of cell membrane potential and electric fields on actin as an excitable system.

ProjectSchematic
By comparing different cell types both on and off ridges, gene expression changes can be correlated with exposure to nanotopographies.

Analysis of Actin Waves using Computer Vision Algorithms

Leonard Campanello, PhD Student, Physics
Matthew J. Hourwitz, PhD Student, Chemistry
with John Fourkas (University of Maryland)

Studying the spatiotemporal dynamics of objects in biological systems can be done in many different ways, including measuring shape dynamics, Particle Image Velocimetry (PIV), and particle-based tracking. However, when attempting to quantify the dynamics of diffuse concentration fields alternative approaches are required. In this project, we utilize a computer vision algorithm called “Optical Flow” to capture and coarse grain the dynamics of amorphous actin intensity fields so that we can measure physical properties such as velocity and force. We use this analysis to determine the effect of surface topographies on actin dynamics, the effect of electric fields on neutrophil migration, and the ways that actin waves in Dictyostelium Discoidium cells change in response to different perturbations.

Optical Flow of Actin
HL60 cell with fluorescently labeled actin, overlaid with optical flow vectors indicating actin dynamics.

3D Reconstruction of T cell Activation Proteins

Leonard Campanello, PhD Student, Physics
with Dr. Maria Traver and Dr. Brian Schaefer (Uniformed Services University) and Dr. Hari Shroff (National Institutes of Health)

Careful regulation of T cell activation is important to ensure that activation signals persist over controlled periods of time. If signals are too short, then the immune system will not properly respond, but if the signals persist for too long, it could result in autoimmune diseases such as Type 1 Diabetes. We are interested in studying the roles of several proteins in regulating T cell signaling, including MALT1 and BCL10. Images are taken on a cutting-edge super-resolution SIM microscope, and the analysis that is done draws on concepts from topology and graph theory.

Structures from Image Analysis

Actin Polymerization in Axon Guidance

Leonard Campanello, PhD Student, Physics
Corbett-Frank, Undergraduate Research Assistant
with Dr. Edward Giniger (National Institutes of Health)

The Abl tyrosine kinase (AKT) signaling network plays an important role in axonogenesis, regulating cell adhesion and actin polymerization. Although individual components within the signaling network are known, the relationship among them remain unresolved. We aim to reveal a subset of this multidimensional relationship: the role of regulated actin polymerization in axon guidance. To probe this relationship, we use growing axons in the wings of Drosophila as a model system. As our aim is to compute a distribution of actin concentration vs. downstream distance along these axons and study the dynamics of this distribution, we must first locate the axon within the 3D image. This project focuses on developing novel filtering algorithms and parallel-medial axis thinning to segment the axons, integrate the actin intensity along the skeleton, and produce the time-dependent actin distribution.

Axon Actin Analysis
Automatically determining the actin-intensity distribution along a growing axon in a Drosophila wing. Top: 2D max-projection of the actin-marked, Z-stacked input image. Bottom: Output of our filtering pipeline: average actin-intensity distribution along the main body of the axon.

Developing Novel Optical Toolsets for Visualizing Cellular Dynamics Across Multiple Scales

Phillip Alvarez, PhD Student, Biophysics
Samira Aghayee, PhD Student, Biophysics
with Dr. Charles Camp (NIST), Dr. Marcus Cicerone (NIST), and Dr. Gabriel Frank (Ben-Gurion University of the Negev)

Through the use of cutting edge optical techniques such as re-scan confocal microscopy, stimulated raman scattering (SRS), and modulation of the quantum properties of light during optical stimulation, new tools being developed in Losert Lab seek to reveal the links between intracellular excitable systems such as the cytoskeleton and action potentials at both single cell and collective scales. The systems currently under development are our hybrid SRS and 2-photon imaging system and our multiscale microscope, the former of which provides for complex optical stimulation and label-free imaging of molecular targets in-vivo and in-vitro, and the latter of which provides a novel approach to simultaneously image at the collective cellular level and at the sub-cellular level beyond the diffraction limit.

BCARS image
Label-free image taken of MDA-MB-231TD cells using Broadband Coherent Anti-Stokes Raman Spectroscopy in collaboration with NIST. Image shows select Raman shifts associated with nucleotides (785 cm-1, 750 cm-1, shown in blue), proteins (1004 cm-1, 1450 cm-1, shown in green), and lipids (2850 cm-1, 1420 cm-1, shown in red).

Tracking-based Motion Correction

Samira Aghayee, Position
with Dr. Patrick Kanold (University of Maryland)

Calcium imaging provides a real time view of neuronal activity with single cell resolution. Neuronal activity is inferred from the dynamics of fluorescence signal coming from the soma membrane. However, as in vivo experiments introduce inevitable jitter to the dataset, any real-time analysis of the network's behavior requires a fast, real-time capable motion correction method. To that end we have introduced a tracking-based motion correction method that reduces the image to a set of a few central positions for the extracted bright neurons in the FOV. The neurons are tracked in time and the image sequence is then corrected to the average position. To compare the performance to the other methods in the field, a simulated dataset with known jitter was generated and the detected jitter is compared to the actual jitter.

Publication: S. Aghayee, D.E. Winkowski, Z. Bowen, E.E. Marshall, M.J. Harrington, P.O. Kanold, and W. Losert, Frontiers in Neural Circuits 2017 [Frontiers]

Motion Correction of Neuron Position

From molecules to cells to organisms: understanding health and disease with multidimensional single-cell methods

Yang Shen, PhD Student, Chemical Physics
with Julian Candia (Research Associate for the Cancer Technology Partnership)

Sophisticated analysis methods for high dimensional single cell datasets are rapidly emerging, with advanced data representations and clustering approaches. However, these are not yet standardized or widely used in the biological fields. In addition, these methods tend to deal with all the measurements simultaneously and hence may suffer from the curse of dimensionality. We are working to build new methods for the automated and unbiased analysis of high-dimensional single cell datasets that is simple, robust, easy to interpret and able to generate testable hypothesis.

Analysis
a) Examples of two types of synthetic point patterns with added-in difference (highlighted with a black circle). Ten replicates were generated for each type of point pattern. CytoBinning (an analysis method we developed) was applied to these point patterns and divided each pattern into 9 regions (labeled B11 to B33). Percentage of cells in each region was calculated. b) Heatmap and hierarchical clustering of aforementioned point patterns based on cell percentage in each region. The two types of point patterns were clustered into distinct groups and the region with built in difference (B32) was correctly identified.

Pattern Recognition in Biodynamics Systems using Machine Learning Algorithms

Qixin Yang, PhD student, Physics

Due to the nature of biological data with large variance, it is important to extract the features that most correlate to the elements of the data that we are most interested in. We are interested in applying machine learning algorithms to facilitate pattern recognition, classification, and prediction in biological systems. We are also looking for new phenotypes to better understand the biodynamics from the machine’s perspective.

Structure of a Neural Network
Convolutional Neural Network (CNN) - Recurrent Neural Network (RNN) architecture designed for classifying biological sequential data.

Maryland Day: Welcome to Cells in Motion!

Lab Outreach

As part of the 2013 Maryland Day event, the Losert Lab prepared a video, "Cells in Motion," which was shown in the inflatable Biomolecular Discovery Dome. We also created an interactive demonstration of our cell tracking software where participants were able to track their motion while playing games such as follow the leader. To watch our video or to learn more, visit our Maryland Day page!

MarylandDay
University of Maryland

Contact

Please contact wlosert @ umd.edu for questions about the Dynamics of Complex Systems lab.