U.S.

Food and Drug Administration

Notice: Archived Document

The content in this document is provided on the FDAs website for reference purposes only. It was current when produced, but is no longer maintained and may be outdated.

How to Quantify Engineered Tissue Structure and Mechanical Behavior

FDA/NIST workshop on In-vitro analyses of cell/scaffold products

Michael Sacks
Department of Bioengineering McGowan Institute for Regenerative Medicine University of Pittsburgh

Tissue Engineering: Role of Biomechanics

Many tissues and organs to be replaced have critical biomechanical functions Tissue Engineering was first coined by Y.C. Fung in 1987 for determining the biomechanical responses of cells and tissues in order to learn how to replace them

Contemporary issues in Biomechanics

Genomic structure Genomic biomechanics Genomic function Molecular structure Molecular biomechanics Molecular function Cell structure Cell biomechanics Cell behavior Tissue structure Tissue biomechanics Tissue function Organ structure Organ biomechanics Organ physiology Human movement Human performance

Biomechanics is the middle name between structure and function

Functional Tissue Engineering*

What are thresholds of force, stress, and strain that the normal tissue must withstand during normal operation? What are their mechanical properties, during both normal and failure conditions? Which properties should be incorporated into TE designs?
*Butler et al., JBME, 2000

Functional Tissue Engineering

When developing implants in culture, how to mechanical factors regulate cell behavior as compared to those experienced in-vivo? Do we have too exactly reproduce every feature of the native tissue to get acceptable levels of physiological function restoration? When evaluating TE repairs, how good is good enough?

Functional Tissue Engineering

1. In-vivo stress/strain histories need to be measured in normal tissues over the physiological range 2. Mechanical properties of the native tissues must be established for sub-failure and failure 3. A subset these mechanical properties must be selected and prioritized 4. Standards must be set when evaluating the repairs/replacements after surgery so as to determine how good is good enough

Primary considerations
In-vitro phase
Enhancement of protein synthesis tissue formation and strength Strategic use of and mechanical/biochemical stimulation

Assessment of In-vivo function

Invasive measures (explant) Non-invasive (primarily image based)

Major scaffold types

Biologically derived
SIS, UBM Decellurized tissues (e.g. aortic valve) Collagen, fibrin, and GAG gels Electrospun biopolymers (collagen)

Synthetic
Wovens and fabrics Gels and foams Non-wovens made from PGA, PLLA Electrospun biodegradable polymers

Mechanical Behavior Driven by physiological functional requirements

a.Stress-strain response non-linearity, rapid transition of stiffness b.Time-dependence
i. viscoelasticity ii.poroelasticity

c. Anisotropy
i. Appropriate knowledge of mechanical properties

d.Dimensionality
i. Uniaxial (tendon) ii.Planar biaxial (valve leaflet) iii.Full 3D (myocardium, cartilage) - No approach available

Mechanical Behavior Driven by physiological functional requirements

1.Major modes
a.Tension b.Compression c.Flexural

2.Meso/macro scale vs. local properties

a.Local
i. AFM, nano-indentation.

b.Larger scales
i. More relevant for physiological function

3.Need to link measures a various scales to make sense of cell and physiological behaviors

Leaflet tri-layered structure

Fibrosa

Spongiosa Ventricularis

Water & Fibroblasts Collagen GAGs Collagen, Elastin Fibrosa ~45% Spongiosa ~35% Ventricularis ~20%

Total thickness ~300-700m

Planar biaxial mechanical properties of the aortic valve leaflet There is more to life than Youngs modulus
radia radial l

60 50
circ rcum umfe fer renti entia al

Circumferential MTM

Radial MTM

Tens io n ( N / m )

40 30 20 10 0
Peak Circumferential Extensibility Peak Radial Extensibility

radial

1.0
circumferential

1.2

1.4

1.6

1.8

Stretch ratio

Structural basis for mechanical behavior

Anisotropy and due to fiber rotations, not stretch
XR

X
C

0:0

Tensi ension on ratios N11:N22

1.0:0.5

1.0:1.0

0.5:1.0

Tissue Engineered Heart Valves (TEHV)

Living autologous cells + bioresorbable scaffolds Potentially have a capacity for growth, self-repair, & resistance to infection Presents opportunity to answer some fundamental bioengineering questions:
How do the scaffold and tissue interact to give rise to overall mechanical properties? How do individual modes of mechanical loading affect tissue development?

Hierarchal Structure of Nonwoven Scaffolds

Primary: Fiber Crimp Tertiary: Discretization by Needling

Secondary: Fiber Orientation

Freed L.E., et al, Bio/technology 1994;12:689-693. Engelmayr, G.C. and Sacks, M.S., J Biomech Eng, 2006

Rational for Cyclic Flexure Bioreactor

Pulse duplicator / Flow loop bioreactor

Used to grow a TEHV for implantation Anatomical geometry Coupled mechanical stimuli

Decompose complex mechanical environment into simple, independent modes of deformation:

Cyclic Flexure Shear Stress Tension Pressure

Why cyclic flexure?

Non-woven scaffolds are not elastomeric Flexure is a mode of deformation innate to heart valves

Aortic valve leaflet dynamic motion

Cyclic Flexure Bioreactor

Engelmayr et al., Biomaterials, 2003, 24(14):2523-32 Engelmayr et al., Biomaterials, 2005, 26(2):175-87

Physiologically Relevant Flexural Testing

Cyclic Flexure Bioreactor Culture Well Three-Point Bending Test

M = EI

Effects of Cyclic Flexure on SMC-Seeded TEHV

Cyclic Flexure

Static

RESView 3D Histology, Resolutions Sciences Corp., Corte Madera CA

A Structural Model for Nonwoven Scaffolds

Flexural Rigidity of RVE Number of fibers per RVE Thickness of RVE

( EI ) RVE =
RVE weight / area

1 12

N f ( E f ) A f t

/2

/2

R ( ) cos d
4

Fiber Orientation Distribution

b Nf =

RVE width

Fiber Cross-Sectional Area Fiber Effective Stiffness

Fiber weight / length

Freeston, W.D., Jr., Platt, M.M., Textile Research Journal 1965;35(1):48-57

Nonlinear Reinforcement Effects of ECM

2000 1800 1600 1400

E (kPa)

1200 1000 800 600 400 200 0 400

450

500

550

600

650

Fiber inter-bond arc length (m)

EECM = 0 EECM > 0

(Ef) = 8896 kPa E = 206 kPa (Ef) ~ 15430 kPa E = 431 kPa

EECM >> 0 (Ef) ~ 55640 kPa E = 1555 kPa

Structural Model Results

Preferred (PD) Fiber Direction

Cross-Preferred (XD) Fiber Direction

Electrospinning setup
Mandrel: 50rpm to 2300rpm or 0.3 m/s to 13.8 m/s
9Aluminum Collection Mandrel

Polymer Poly (ester urethane) urea (PEUU)

ES-PEUU microstructure

0.0 m/s

0.3 m/s

1.5 m/s

3.0 m/s

4.5 m/s

9.0 m/s

13.8 m/s

Why ES-PEUU scaffolds ?

Electrospun PEUU scaffolds exhibit
wide range of mechanical compliance and anisotropy mechanical properties very similar to native tissue

Mechanical analysis
Increasing Mandrel Velocity Membrane Tension (N/m) 90 60 30 0 90 60 30 0 1.0 1.1 1.2 1.3 (stretch) 1.4 1.5 1.6
Cross-preferred Direction

0.0 m/s 0.3 m/s 1.5 m/s 3.0 m/s 4.5 m/s 9.0 m/s 13.8 m/s

Preferred Direction

Increasing Mandrel Velocity

Equibiaxial stress-stretch results

Comparison to native pulmonary valve

Model formulation
Stress-stretch relations

P11 =

Sf [Ef ()] R() (F11 cos2 + F12 sin cos )d

2

P22 =

Sf [Ef ()] R() (F22 sin2 + F21 sin cos )d

2

Effective fiber properties

Effective fiber structure

Kinematic terms based on experimental strains

Structural model fit

700 1st Piola Kirchhoff Stress (kPa)

700 1st Piola Kirchhoff Stress (kPa)

600 500 400 300 200 100

Preferred Direction

600 500 400 300 200 100

Cross-Preferred Direction

Experimental Data Structural Model Fit

0.0 m/s or Random

1.30

Experimental Data Structural Model Fit

0 0.95

1.00

1.05

1.10

1.15

1.20

1.25

0 0.95

1.00

1.05

1.10

1.15

1.20

1.25

1.30

600
1st Piola Kirchhoff Stress (kPa) 600

Preferred Direction

1st Piola Kirchhoff Stress (kPa)

500 400 300 200 100

Cross-Preferred Direction

400

200
Experim ental Data Structural Model Fit

4.5 m/s Or 750 rpm

1.06

Experimental Data Structural Model Fit

0 0.98

0.99

1.00

1.01

1.02

1.03

1.04

1.05

0 0.95 1.00 1.05 1.10 1.15 1.20 1.25 1.30 1.35

Effective fiber stress-stretch

350 300 250

Increasing Mandrel Velocity

Tf (kPa)

200 150 100 50 0 1.00

random 0.3 m/s 1.5 m/s 3.0 m/s 4.5 m/s 9.0 m/s 13.8 m/s

1.02

1.04

1.06

1.08

From this model we can

1. Obtain true fiber (polymer) moduli as opposed to effective fiber stress-strain response using exponential model used previously 2. Separate structural effects (e.g. orientation) from changes in fiber material properties 3. Allow derivation of true fiber (material) moduli independent of micro-structural features 4. Practical uses:
Guiding scaffold design for tissue or cell specific applications optimizing in-vitro conditioning regimes to produce viable tissues for implantation.

Scaffold physical characterization-Structure

many methods are available
Porosity and pore geometry Focus on fibrous architecture as this dictates both bulk properties and local cellular deformations
SALS for both native and non-wovens. EM and CLSM fiber alignment image analysis

Cellular deformations and it relation to local and global fiber architecture.

Native tissue as the functional endpoint CLSM of cell micro-integrated scaffolds

Laser Light Scattering

Fiber axis Scattered light

Scattering pattern

Incident light

Tissue specimen Light is scattered perpendicular to fiber axis HeNe Laser, = 632.8 nm

Laser Light Diffraction

In connective tissues, one slit is the fiber and the other the gap between fibers Since the gap must follow the fiber geometry, it turns out that this distinction is not necessary

fiber gap fiber gap fiber

Angular Fiber Distribution from SALS

Meridian
Scattered light intensity (A/D units)

Fiber preferred direction

120

s
100

80

60

Equator

40

20

OI
0 -45 -30 -15 0 15 30 45 60 75 90 105 120

(degrees)

Native valvular tissue

OI (deg)

SALS was used to compare the changes in fiber distributions between pressurefixed aortic valve cusps (bottom row) and non-pressure fixed (top row). The changes in crimp due to the two preparations are on the right. The SALS data (center) shows a much higher alignment in the pressure-fixed cusp.

SIS Multiple Fiber Populations

70 65 60 55 50 45 40 35 30 25
80 Intensity (a/d units) 60 40 20 0 -80 -60 -40 -20 0 20 40 60 80 (degrees)

Test #343
Intensity (a/d units)

80

Test #910 c1 c2 2

c1 1

c2 2

60 40 20 0

-80 -60 -40 -20 0 20 40 60 80 (degrees)

SALS can also indicate the presence of multiple fiber populations, which can then be deconstructed using mathematical techniques to investigate the results of multiple fiber populations on mechanical performance.

Engineered Biomaterials
1.0

Dermagraft

Dermagraft data Eqn 1 fit Isolated mesh data

Normalized intensity

0.8

Collagen Isolated mesh

0.6

0.4

0.2

2 3
0 30 60 90

1 4
120 150 180

0.0

(degrees)

SALS can be used to evaluate the structural properties of composite biomaterials such as the Dermagraft (Advanced Tissue Sciences). This material is composed of a biodegradable mesh embedded in a collagen matrix. Both the collagen and mesh components of the fiber distribution are observed in the SALS signal (right).

R(): Normalized Fiber orientation distribution of non-wovens

0.05 50:50 PGA/PLLA PGA PLLA

0.04

0.03

0.02

0.01

2 1 + exp 2 y0 2 2 R ( ) = /2 2 1 + exp 2 d / 2 2 y0 2

R()

0.00 -90 -60 -30 0 30 60 90

(degrees)

Material PGA PLLA 50:50

Thickness (m) 948 28 1153 37 889 6

Bulk Density (mg/cm3) 69.0 61.9 61.75

Mean Fiber Orientation Distribution, R() Normalized Gaussian Model (= 0) R2 (degrees) yo 31.66 25000 0.9951 33.79 20000 0.9986 32.97 15000 0.9987

SALS Tendon tissue engineering*

*Nirmalanandhan VS, Rao M, Sacks MS, Haridas B, Butler DL., JB 2007

Global deformations vs. local fiber response

Illustration of multi-scale characteristics encountered when relating global deformations to local fiber responses
XD PD

PD
Spinning Mandrel

XD

Fiber orientation and tortuosity tracking

a. b.

1.0 0.8 0.6 0.4 0.2 0.0 -100 -80 -60 -40 -20 0 20 40 60 80 100 Degrees

Custom image analysis to quantify (a) orientation and (b) tortuosity

R()

Fiber architecture analysis*

Tracking fiber splay
1.6 1.4 1.2 1.0 R() 0.8 0.6 0.4 0.2 0.0 -100 -80 Increasing Mandrel Velocity
0 m/s 0.3 m/s 1.5 m/s 3.0 m/s 4.5 m/s 9.0 m/s 13.8 m/s

-60

-40

-20

20

40

60

80

100

Fiber splay results for all specimens

Degrees

*Courtney et al., Biomaterials, 2006

Structural uniformity

Functional Tissue Engineering Effects of changes in tissue formation with time

a.Effects of tissue formation.
i. Physical stimulation to enhance tissue generation. ii.Methods to assess effects of tissue formation.

b.Scaffold degradation
i. Mass changes ii.Surface vs. bulk erosion iii.Stress-transfer considerations.

Mechanical training paradigm

INPUTS 1.
Controlled cell deformation

OUTPUTS 1.
Phenotype changes

2.
(%)

Loading time and wave form

Mechano-dependent, phenotypic/biosynthetic response

Biosynthetic levels

2.
Robust ECM formation Scaffold degradation

3. 4.

# of cycles Addition of: growth factors ascorbic acid

Related studies
Relating the microenvironment experienced by a cell in response to global tissue deformation is a reoccurring question
Cellular deformation influences biosynthetic activity

Mow et al. Chondrocyte deformation and local tissue strain in articular cartilage In recent studies, Huang et al. investigated the response of aortic valve interstitial cells (AVICs) with increasing transvalular pressure Cell nuclear aspect ratio was used to measure cell deformation
Huang,

et al. Effects pf transvalvular pressure on the aortic valve interstitial cell nuclear aspect ratio. JBME. In-press

VIC deformations within HV tissues

0 1

90

Collagen alignment-VIC aspect ratio

5.5 6.0 5.0 5.5 60 mmHg

AVIC NAR VIC nucleus aspect ratio

5.0 4.5 4.5 4.0 4.0

AVIC compression Cell compression

90 mmHg

60 mmHg

TVP ~4 mmHg TVP 0 mmHg

3.5 3.5 3.0 3.0

2.5 2.5 2.0 2.0 1.5

2 mmHgfiber Collagen alignment 4 mmHg 1 mmHg

2 mmHg

Collagen fiber straightening 4 mmHg

1 mmHg
0 mmHg

0 mmHg

(b)
52 0.60 54 56 0.65 58 0.70

1.5 1.0 38 0.30

40 0.35

420.40 44 0.45 46

More aligned

Orientation Index (deg) Normalized orientation index

48 0.50

50 0.55

TVP ~60 mmHg Compaction effects dominate

Less aligned

Mechanical stimulation of heart valve tissues1

aorta NC LC RC
3.5

Ratio of Hsp47 from day 0

3.0 2.5 2.0 1.5 1.0 0.5

control

14 days left ventricle

7 days

Null Tension TGF Tension+TGF

14

Days

1Merryman

et al. Cardiovascular Pathology, in-press

Effects of Cyclic Flexure on SMC-Seeded TEHV

Cyclic Flexure

Static

RESView 3D Histology, Resolutions Sciences Corp., Corte Madera, CA51

Key Results from Cyclic Flexure Studies

Trend of Increased effective stiffness with cyclic flexure compared with static 63% increase in collagen concentration with cyclic flexure

Mechanical Stimulation
Effective stiffness is highly dependent on collagen concentration
2000

Cyclic flexure can homogenize the transmural cell and ECM distribution
0.16 0.14 Static Flex

Structural Mechanics

1800 1600 1400

Normalized Cell Count

2000 2500

t = 30 Hours t = 3-Week Static t = 3-Week Flex t = 9-Week TEHV Linear Regression

0.12 0.10 0.08 0.06 0.04 0.02 0.00 -0.5

E (kPa)

1200 1000 800 600 400 200 0 0 500 1000 1500 R2 = 0.996

-0.4

-0.3

-0.2

-0.1

0.0

0.1

0.2

0.3

0.4

0.5

Collagen Concentration (g/g wet weight)

Normalized Transmural Thickness

Engineered Heart Valve Tissue

PGA scaffold TEHV scaffold after 18 day dynamic incubation

Meso-Scale Model for Nonwoven-ECM

( EI ) RVE =

t/2

t / 2

E ( y ) y dydx

ECM-Coupling Parameter ECM Effective Stiffness Scaffold Effective Stiffness [Collagen]

E ( y ) = R EECM ( y ) + Es

EECM ( y ) = C ( y ) Bc Ec
Normalized Transmural Collagen Concentration Collagen Specific Stiffness (i.e., stiffness/quantity)

( EI ) RVE =

2 t / 2 R ( C ( y) Bc Ec ) + Es y dydx t/2

Engelmayr and Sacks, Biomech Model Mechanobiol, 2006, in preparation

C(y): Normalized Transmural Collagen Concentration Distribution

Fluorescence Microscopy of Picro-Sirius Red Stained Sections Image Analysis for Normalized Fluorescence Intensity Distribution

Mean Normalized Fluoresensce Intensity

0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00 0 200 400 600 800 1000 1200 1400 1600

Normalized Collagen Concentration, C(y)

4 Flex Static 3

-0.4

-0.2

0.0

0.2

0.4

Collagen Concentration (g/g wet weight)

Normalized Transmural Thickness, y

Dolber, P.C. and Spach, M.S., J Histochem Cytochem, 1993, 41(3):465-9.

R: Nonwoven-ECM Coupling Parameter

Physical Models of ECM and TEHV: Polyacrylamide (Pam) gel and Pam gel-infiltrated nonwoven scaffold
3500 3000 2500

R is determined from linear region of E versus EECM plot

E (kPa)

2000 1500 1000 500 0 0 200 400 600 800

R = 7.444 kPa/kPa

Rule-of-mixtures

EECM (kPa)

E = R EECM + Es

E = EECM vECM + Es

EECM(y): Transmural ECM Effective Stiffness

Experimental Group Static Flex E (kPa) 748 130 978 228 Bc (g/g wet weight) 893 133 546 111 C(y) =ay2+c a c 9.946 0.173 4.009 0.667 c (kPa/( g/g wet weight) 0.0789 0.0904

180 160 140 120 Flex Static

EECM (kPa)

100 80 60 40 20 0 -0.4 -0.2 0.0 0.2 0.4

Normalized Transmural Thickness, y

Next steps: Scale up

Local cellular deformations need to be controlled at the macro-level in an intact valve Need to balance need for controlled biomechanical stimulation with other valve design requirements

Physiological flow loop bioreactor1

Hildebrand et al. Annals of Biomedical Engineering 2004

Physiological flow loop bioreactor

Stent Belly Free edge

FE simulations of leaflet principal strain (quasi-static loading)

ANSYS 8.1

ANSYS 8.1

-.566E-06 .031624

.063249 .094873

.126498 .158122

.189747 .221372

.252996 .284621

-.175803 .080168

.33614 .592111

.848083 1.104

1.36 1.616

1.872 2.128

Isotropic fiber distribution

2300 rpm fiber distribution

Ongoing issues and future trends

Lots of techniques and approaches what is the correct approach?
Driven by functional understanding and application.

There is more to life than Youngs modulus what do you measure? Biomechanical studies usually require large specimens and large number of specimens due to variability cost/benefit.

Ongoing issues and future trends

a.Need for non-destructive simultaneous cell/tissue imaging during in-vitro incubation and in-vivo development.
i. Optical methods ii.US iii.MRI

b.Need for standardization of approaches ASTM? c. Need for low cost, high throughput, physiologically meaningful tests.
i. Role of commercial sector.

Acknowledgments!
Graduate students: George Engelmayr, Dan Hildebrand, David Merryman, Todd Courtney, David Schmidt, John Stella, John Stankus, Nick Amoroso, Chad Eckert Research Faculty: Jun Liao, Jinjuan Guan, Yi Hong, David Schmidt, Sharan Ramaswamy Collaborators: John E. Mayer, Jr., Frederick J. Schoen, Elena Rabkin, Richard Hopkins, and William Wagner NHLBI R01s: HL-68816 and HL17649 NIBIB T32 Biomechanics in Regenerative Medicine