Tissue Biophysics and Biomimetics

Peter J. Basser, PhD, Head, Section on Tissue Biophysics and Biomimetics
Ferenc Horkay, PhD, Staff Scientist
Carlo Pierpaoli, MD, PhD, Staff Scientist
Lin-Ching Chang, PhD, Postdoctoral Fellow
Cheng Guan Koay, PhD, Postdoctoral Fellow
David Lin, PhD, Postdoctoral Fellow
Joelle Sarlls, PhD, Postdoctoral Fellow
Uri Nevo, PhD, Visiting Fellow
Evren Özarslan, PhD, Visiting Fellow
Iren Horkayne-Szakaly, MD, Volunteer
Michal Komlosh, PhD, Volunteer
Candida Silva, PhD, Volunteer
Ichiji Tasaki, MD, PhD, Volunteer¹
Lindsay Walker, MS, Volunteer

We try to understand fundamental relationships between function and structure in soft tissues, “engineered” tissue constructs, and tissue analogues—how microstructure, hierarchical organization, composition, and material properties of tissue affect biological function and dysfunction. To investigate biological and physical model systems at different time and length scales, we make physical measurements in tandem with analytical and computational models. Primarily, we use water to probe both equilibrium and dynamic interactions among tissue constituents over a wide range of time and length scales. To determine the equilibrium osmo-mechanical properties of well-defined model systems, we vary water content or ionic composition. To probe tissue structure and dynamics, we employ small-angle X-ray scattering (SAXS), small-angle neutron scattering (SANS), static light scattering (SLS), dynamic light scattering (DLS), and nuclear magnetic resonance (NMR) relaxometry. We use mathematical models to understand how changes in tissue microstructure and physical properties affect essential transport processes (e.g., mass, charge, and momentum). The most direct non-invasive method for characterizing these transport processes in vivo is magnetic resonance imaging (MRI), which can characterize normal tissue microstructure and follow its changes in development, degeneration, and aging. Another goal is to translate our new quantitative methodologies from bench to bedside.

Virtual in vivo tissue biopsy

Basser, Pierpaoli, Komlosh, Nevo, Özarslan; in collaboration with Assaf, Cohen, Freidlin, Jarisch, Pajevic, Pikalov

We continue to develop novel magnetic resonance (MR) methods that allow us to probe tissue microstructure and diagnose neurological and developmental disorders in vivo. Diffusion Tensor MRI (DT-MRI or DTI) is the most mature such technology that we have developed. With it, we measure a diffusion tensor of water, D, on a pixel-by-pixel basis within tissue. Information derived from D includes the local fiber-tract orientation, the mean-squared distance that water molecules diffuse in any given direction, the orientationally averaged mean diffusivity, and other intrinsic scalar (invariant) quantities that are independent of the laboratory coordinate system. The scalar parameters behave like quantitative histological “stains” but are “developed” without exogenous contrast agents or dyes. The bulk or orientationally averaged diffusivity is the most successful imaging parameter proposed to date for identifying ischemic tissue regions in the brain during acute stroke. Furthermore, measures of diffusion anisotropy are useful in identifying white matter degeneration (e.g., Wallerian degeneration) associated with chronic stroke. DTI also provides new information about early developmental changes in cortical gray and white matter that cannot be detected with other imaging methods.

We have also pioneered the use of color maps to encode nerve fiber orientation, allowing us to identify the main association, projection, and commissural white matter pathways in the brain and even to differentiate white matter pathways with similar structure and composition but distinct spatial orientations. To assess anatomical connectivity between different functional brain regions, we developed DTI fiber tractography to track nerve fiber trajectories by continuously following the direction along which the apparent diffusivity is at a maximum. While the reliability of DTI tractography is high in coherently organized primary white matter pathways, artifactual tracts may be generated when following white matter pathways with a more complex underlying fiber topology (e.g., pathways that “kiss,” cross, merge, or branch). To assess the reliability of individual computed tracts, we have used Point-wise Assessment of Streamline Tractography Attributes (PASTA). We are using other MR data and more detailed models of water diffusion in tissue to compensate for artifacts.

We have proposed several advanced in vivo MR methods to measure fine microstructural features of nerve fascicles that previously could be measured only by using laborious histological methods. One approach is CHARMED (Composite Hindered And Restricted Model of Diffusion), a combined experimental and modeling framework that allows us to describe water diffusion in white matter. More specifically, CHARMED permits us to distinguish between water diffusing within the intraaxonal spaces and extra-axonal compartments, thereby improving the angular resolution in tract tracing and resolution of fibers that cross. Recently, we extended CHARMED (and dubbed it AxCaliber) in order to measure the axon diameter distribution within large fascicles in vivo.

While gray matter appears featureless in DTI, its microarchitecture is rich and varied. We have been working toward developing a non-invasive, in vivo method to perform Brodmann parcellation of the cerebral cortex. To this end, we are developing advanced MR sequences to probe correlations between displacements of water molecules within brain tissue and various gray matter compartments at microscopic resolution as well as sophisticated mathematical models describing water displacements in those compartments; we use the latter to infer key microstructural and morphological features.

To permit analysis of novel multidimensional data sets, clinical and biological applications of DTI (and of other novel displacement MRI methods) generally require new mathematical, statistical, and image-sciences concepts and tools. We have developed algorithms for continuous, smooth approximation to the discrete, noisy, measured DT field data; the algorithms reduce noise and enable us to follow fibers more reliably. We recently derived a new Gaussian distribution for tensor-valued random variables that we used in designing optimal DTI experiments. In addition, we have developed non-parametric empirical (e.g., Bootstrap) methods for determining the statistical distribution of DT-derived quantities in order to study, for example, the inherent variability and reliability of white matter fiber tract trajectories. These parametric and non-parametric statistical methods enable us to apply powerful statistical hypothesis tests to a wide range of important biological and clinical questions that previously could be examined only with ad hoc methods. More recently, we have been developing a variety of economical empirical methods to describe the effect of noise on DTI data, permitting the assessment of data variability with a clinically feasible MRI acquisition.

Basser PJ, Pajevic S. Spectral decomposition of a 4th-order covariance tensor: applications to diffusion tensor MRI. Signal Processing 2007;87:220-236.
Freidlin RZ, Özarslan E, Komlosh ME, Chang L-C, Koay CG, Jones DK, Basser PJ. Parsimonious model selection for tissue segmentation and classification applications: a study using simulated and experimental DTI data. IEEE T Med Imaging 2007;26:1576-1584.
Jian B, Vemuri BC, Özarslan E, Carney PR, Mareci TH. A novel tensor distribution model for diffusion weighted MR signal. Neuroimage 2007;37:164-176.
Komlosh ME, Horkay F, Freidlin RZ, Nevo U, Assaf Y, Basser PJ. Detection of microscopic anisotropy in gray matter and in a novel tissue phantom using double Pulsed Gradient Spin Echo MR. J Cardiovasc Magn Reson 2007;189:38-45.
Shepherd TM, Özarslan E, Yachnis AT, King MA, Blackband SJ. Diffusion tensor microscopy indicates the cytoarchitectural basis for diffusion anisotropy in the human hippocampus. Am J Neuroradiol 2007;28:958-964.

MRI study of normal brain development

Pierpaoli, Basser, Chang, Sarlls, Walker, Koay, Özarslan; in collaboration with The Brain Development Cooperative Group

Four NIH Institutes (NICHD, NIMH, NINDS, and NIDA) have jointly sponsored a multicenter study to advance our understanding of brain development in typical, healthy children and adolescents. The study has enrolled approximately 500 children, ranging from infancy to young adulthood, who have been seen at different times over a six-year period (2001–2007) at several clinical centers around the United States. The group has acquired MR images of the brain and will relate the imaging findings to the results of standardized neuropsychological tests.

Our role in this project is to serve as a diffusion tensor data processing center (DPC). We process and analyze all DTI data that the various clinical centers acquire during the course of the study. During the past several years, we have been developing and implementing a novel DTI data processing pipeline. Specifically, we have developed procedures for sorting, displaying, and co-registering the raw diffusion-weighted images from which the DTI data are computed. The image registration procedure removes the effects of subject motion and eddy current distortion and aligns the images to a given template with only one interpolation step, ensuring minimal loss of data quality. We also proposed a new strategy for robust estimation of the diffusion tensor and the quantities derived from it. We undertook the task of registering DTI data to other structural MRI data in the database (which is public) by using rigid body and linear (affine) transformations. More recently, we addressed the issue of correcting residual image distortion originating from the Echo-Planar Image (EPI) acquisition used for DTI.

In 2005, the project received Neuroscience Blueprint funds to expand the DTI portion of the study. The original DTI scan was limited to a 10-minute protocol performed on those subjects able to tolerate additional time in the scanner following the collection of the core structural MRI data. The study’s expansion is increasing the quality and quantity of data collected through longer or additional scanning sessions for older subjects and through the recruitment of new subjects and scanning sessions for the younger cohort. Our group has designed the expanded DTI protocol and coordinated its implementation at each acquisition site. We completed data collection in 2007 and are now receiving raw diffusion data and structural MRI targets for processing through our pipeline. We plan to produce processed diffusion tensor data to be made available to the scientific community approximately one year after receiving all the raw data.

Almli CR, Rivkin MJ, McKinstry RC, Brain Development Cooperative Group (Pierpaoli C, member). The NIH MRI study of normal brain development (Objective-2): newborns, infants, toddlers, and preschoolers. Neuroimage 2007;35:308-325.
Chang LC, Koay CG, Pierpaoli C, Basser PJ. Variance of estimated DTI-derived parameters via first-order perturbation methods. Magn Reson Med 2007;57:141-149.
Irfanoglu MO, Machiraju R, Sammet S, Pierpaoli C, Knopp MV. Automatic deformable diffusion tensor registration for fiber population analysis. MICCAI 2008, Part II, LNCS 2008;5242:1014-1022.
Koay CG, Sarlls JE, Özarslan E. Three-dimensional analytical magnetic resonance imaging phantom in the Fourier domain. Magn Reson Med 2007;58:430-436. [Software download].
Waber DP, De Moor C, Forbes PW, Almli CR, Botteron KN, Leonard G, Milovan D, Paus T, Rumsey J, Brain Development Cooperative Group. The NIH MRI study of normal brain development: performance of a population based sample of healthy children aged 6 to 18 years on a neuropsychological battery. J Int Neuropsychol Soc 2007;13:1-18.

Biopolymer physics: water-ion-biopolymer interactions

Horkay, Lin, Basser; in collaboration with Geissler, Hecht, Tasaki

To help us understand the nature of physical/chemical interactions within biomolecules and among biomolecular assemblies, we developed an experimental approach to study assembly structure (morphology) and thermodynamic properties as a function of length scale (i.e., spatial resolution). The methodology combines macroscopic (coarse length–scale) osmotic swelling pressure measurements and various high-resolution scattering experiments. Macroscopic swelling pressure measurements provide information about the overall thermodynamic response, whereas SANS, SLS, and DLS collectively provide information about the sizes of structural elements and their respective contribution to macroscopic osmotic properties. Combining these measurements allows us to (1) separate the scattering intensity caused by thermodynamic concentration fluctuations from large static superstructures (e.g., aggregates) and (2) determine the length scales of fluctuations that give rise to macroscopic thermodynamic properties.

Moreover, in solutions of large charged polymers, counter-ions form a cloud surrounding the polymer chain. With increasing salt concentration, a fraction of these ions adsorb onto the macroion. For monovalent ions (such as sodium), their distribution can be approximated by the Poisson-Boltzmann theory; however, the theory fails to describe the counter-ion atmosphere in the presence of multivalent counter-ions (such as calcium), whose effect on chain conformation remains poorly understood. Knowledge of the counter-ion distribution is therefore essential to understanding the behavior of biomolecules in the physiological milieu. Recently, we developed a method based on measurement of the X-ray scattering signal in the vicinity of the absorption edge of the counterions that allows us to determine the distribution of ions around charged biopolymer molecules.

Specifically, we have applied the above multiscale approach to the interactions of multivalent cations, particularly calcium, on the structure and morphology of various biomolecules. Divalent cations are ubiquitous in the biological milieu, yet existing theories do not adequately explain their effect on and interactions with charged polymers or biomolecules. Moreover, experiments to study the interactions are difficult to perform, particularly in solution, because multivalent cations above a low concentration threshold generally cause phase separation or precipitation of charged molecules. Given that macroscopic phase separation does not occur in cross-linked gels, we overcame this limitation by cross-linking our biopolymers, thereby extending the range of ion concentrations over which the system remains stable and may be studied. We have applied the new method in pilot studies of DNA and of cross-linked gels of the model synthetic polymer polyacrylic acid. Concentrated DNA solutions and novel DNA gels have never before been investigated with SANS in conjunction with osmotic measurements. Our method also provides a unique framework for analyzing the osmotic and scattering behavior of other biomolecular systems.

We also investigated cross-linked gels of the biopolymer hyaluronic acid (HA) to determine the size of the structural elements that contribute to osmotic concentration fluctuations. We combined SAXS and SANS to estimate the osmotic modulus of HA solutions in the presence of monovalent and divalent counterions. We also investigated the diffusion processes in biopolymer solutions, using DLS to determine the osmotic modulus independently from the dynamic response.

Geissler E, Hecht A-M, Horkay F. Scaling equations for a bipolymer in salt solution. Phys Rev Lett 2007;9:267801(1-4).
Horkay F, Basser PJ, Hecht A-M, Geissler E. Comparative study of scattering and osmotic properties of synthetic and biopolymer gels. Macromol Symp 2007;256:80-87.
Horkay F, Hammouda B. Small-angle neutron scattering from typical synthetic and biopolymer solutions. Colloid Polym Sci 2008;286:611-620.
Michelman-Ribeiro A, Horkay F, Nossal R, Boukari H. Probe diffusion in aqueous poly(vinyl alcohol) solutions studied by fluorescence correlation spectroscopy. Biomacromolecules 2007;8:1595-1600.

Functional properties of extracellular matrix

Horkay, Lin, Silva, Basser; in collaboration with Dimitriadis

The swelling behavior of cartilage is exquisitely sensitive to biochemical and microstructural changes during development, disease, degeneration, and aging. Given that primarily the osmotic properties of its constituents allow cartilage to resist applied loads and express or imbibe fluid, understanding the mechanisms affecting cartilage hydration is also essential to predicting cartilage’s load-bearing and lubricating ability.

Previously, we showed that controlled hydration or swelling of cartilage is a means of determining functional properties of cartilage’s extracellular matrix (ECM), thus permitting an independent measurement of important physical/chemical properties of the collagen network and proteoglycans (PG) within the ECM. This approach involves (1) modeling the ECM as a composite material consisting of two distinct phases—a collagen network and a proteoglycan solution trapped within it; (2) applying various known levels of equilibrium osmotic stress; and (3) using physical-chemical principles and additional experiments to determine independently a pressure-volume relationship for both the PG and collagen phases. In pilot studies, we used this approach to determine pressurevolume curves for the collagen network and PG phases in native and trypsin-treated normal human cartilage specimens as well as in cartilage specimens from osteoarthritic (OA) joints. In both normal and trypsin-treated specimens, collagen network stiffness appeared unchanged, whereas collagen network stiffness was reduced in OA specimens. Our findings highlighted the role of the collagen network in limiting normal cartilage hydration and ensuring a high PG concentration in the matrix, which are both essential for effective load bearing in cartilage and lubrication but are lost in OA. The data also suggest that the loss of collagen network stiffness, rather than the loss or modification of PGs, may be the incipient event leading to the subsequent disintegration of cartilage observed in OA.

One shortcoming of this approach, however, was that it took many days to analyze a single cartilage specimen, making the technique unsuitable for routine pathological analysis or tissue engineering applications. Furthermore, a significant amount of tissue was required to perform these osmotic titration experiments. Therefore, we developed a new tissue micro-osmometer to perform such experiments practically and rapidly. The instrument can measure minute amounts of water absorbed by small tissue samples (less than 1 µg) as a function of the equilibrium activity (pressure) of the surrounding water vapor. A quartz crystal detects the water uptake of a specimen attached to the crystal’s surface. The high sensitivity of the crystal’s resonance frequency to small changes in the amount of adsorbed water allows us to measure precisely the mass of water taken up by the tissue specimen. Varying the equilibrium vapor pressure surrounding the specimen induces controlled changes in the tissue layer’s osmotic pressure.

To illustrate the applicability of the new apparatus, we measured the swelling pressure of a tissue-engineered cartilage specimen. We also used atomic force microscopy (AFM) in tandem to determine and map the tissue’s local-mechanical properties and biochemical analysis to determine the concentration of the main biopolymer components. In addition, we study the contribution from individual components to the tissue’s osmotic and mechanical properties. Our current research focuses on aggrecan, which is the most abundant cartilage proteoglycan and is most responsible for the tissue’s load-bearing ability. We study the static properties of aggrecan and aggrecan/hyaluronic acid aggregates by SLS, SANS, and SAXS and by osmotic pressure measurements and we probe the dynamics by DLS and neutron spin-echo measurements.

The micro-osmometer will eventually permit us to obtain a simultaneous profile of the osmotic compressibility or stiffness of several cartilage specimens as a function of depth from the articular surface to the cartilage/bone interface, to quantify the contributions of individual components of ECM (e.g., aggrecan, hyaluronic acid, and collagen) to total osmotic pressure, and to assess the osmotic compatibility and mechanical integrity of developing tissues and tissue-engineered cartilage (or ECM) for improved integration following implantation in reconstructive or restorative surgical procedures.

We recently developed an AFM technique for mapping the local elastic properties of tissues and successfully addressed many of the issues that previously hindered use of AFM as a high-throughput probe of inhomogeneous samples, particularly of biological tissues. The technique uses the precise scanning capabilities of a commercial AFM to generate large volumes of compliance data and automatically extracts the relevant elastic properties. Used in conjunction with results from microosmometry and biochemical analysis, AFM will allow us to map spatial variations in the load-bearing capacity of cartilage specimens.

Bencherif SA, Srinivasan A, Horkay F, Hollinger JO, Matyjaszewski K, Washburn NR. Influence of the degree of methacrylation on hyaluronic acid hydrogels properties. Biomaterials 2008;29:1739-1749.
Horkay F, Basser PJ. Insensitivity to salt of assembly of a rigid biopolymer aggrecan. Phys Rev Lett 2008;101:068301(1-4).
Lin DC, Dimitriadis EK, Horkay F. Elasticity of rubber-like materials measured by AFM nanoindentation. eXPRESS Polym Lett 2007;1:576-584.
Lin DC, Dimitriadis EK, Horkay F. Robust strategies for automated AFM force curve analysis–I. Non-adhesive indentation of soft, inhomogeneous materials. J Biomech Eng (Transactions of the ASME) 2007;129:430-440.
Lin DC, Dimitriadis EK, Horkay F. Robust strategies for automated AFM force curve analysis–II. Adhesion-influenced indentation of soft, elastic materials. J Biomech Eng (Transactions of the ASME) 2007;129:904-912.

¹Scientist Emeritus (NIMH)

Collaborators

Yaniv Assaf, PhD, Tel Aviv University, Tel Aviv, Israel
Alan Barnett, PhD, Clinical Brain Disorders Branch, NIMH, Bethesda, MD
Yoram Cohen, PhD, Tel Aviv University, Tel Aviv, Israel
Emilios Dimitriadis, PhD, Division of Bioengineering and Physical Science, ORS, NIH, Bethesda, MD
Raisa Freidlin, MS, Computational Bioscience and Engineering Laboratory, CIT, NIH, Bethesda, MD
Erik Geissler, PhD, CNRS, Université Joseph Fourier de Grenoble, Grenoble, France
Anne-Marie Hecht, PhD, CNRS, Université Joseph Fourier de Grenoble, Grenoble, France
Wolfram Jarisch, PhD, LifeStar LLC, Potomac, MD
Stefano Marenco, MD, Clinical Brain Disorders Branch, NIMH, Bethesda, MD
Pedro Miranda, PhD, Universidade de Lisboa, Lisbon, Portugal
Sinisa Pajevic, PhD, Mathematical and Statistical Computing Laboratory, CIT, NIH, Bethesda, MD
Valery Pikalov, PhD, Institute of Theoretical and Applied Mechanics of the Russian Academy of Sciences, Novosibirsk, Russia
The Brain Development Cooperative Group

For further information, contact pjbasser@helix.nih.gov or visit http://stbb.nichd.nih.gov.