| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* Program in Molecular and Integrative Physiological Sciences, Department of Environmental Health, Harvard School of Public Health, Boston, Massachusetts, USA; and
Division of Engineering and Applied Sciences and Department of Physics, Harvard University, Cambridge, Massachusetts, USA
1Correspondence: HSPH-Environmental Health, 665 Huntington Ave., Boston, MA 02115, USA. E-mail: dtschump{at}hsph.harvard.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: compliance • extracellular matrix • remodeling • myofibroblast • fibrosis
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. The ECM also provides a rich signaling environment, contributing biochemical and mechanical cues important to morphogenesis, wound healing, and disease processes (1
2
3)
. Thus the mechanical properties of a tissue’s ECM are optimized for proper organ function, and these same properties strongly influence the functional behaviors of resident cells (4
5
6
7)
. When pathophysiological changes in tissue mechanics develop (2
, 3
, 5
, 6)
, therapeutic approaches require a mechanistic understanding of the signals and cellular processes that regulate matrix mechanics (2)
. Although sophisticated tools are available to study the global and local changes in matrix architecture mediated by cell-matrix interactions (2
, 8
9
10)
, fewer options exist to quickly and efficiently measure cell-mediated changes in matrix mechanical properties. The objectives of this study are to validate a new method for measuring the mechanics of a simplified matrix (type I collagen gels) undergoing cell-mediated remodeling and to demonstrate the assay’s unique screening capabilities by measuring changes in matrix stiffness driven by a panel of cytokines implicated in fibrotic disorders.
Previous investigations of cell-mediated matrix remodeling have relied extensively on fibroblasts, the primary matrix regulatory cell, embedded within 3-dimensional scaffolds of purified reconstituted extracellular matrix proteins. Such models have underscored the distinct ways cells interact with 3-dimensional matrices as opposed to traditional 2-dimensional tissue culture surfaces (11)
. The 3-dimensional models have been instrumental in the discovery and mechanistic understanding of cell-mediated matrix compaction, and the influence of an array of soluble stimuli on this process (reviewed by Grinnell, ref. 10
). Moreover, these 3-dimensional models have demonstrated important roles for matrix dimensionality (12)
and mechanics (13)
in regulating gene expression, particularly expression of matrix constituents, emphasizing the potential for cells to sense and change the mechanics of their environment. However, measuring changes in gene and protein expression, even those strongly implicated in matrix biology, is likely to be only a rough proxy for measuring cell-mediated changes in matrix mechanical properties.
To directly assess the mechanical properties of extracellular matrix preparations, traditional engineering methods such as rotational shear (14
, 15)
, confined compression (16)
, and uniaxial tension (17
, 18)
testing have been used. While these methods allow comprehensive testing of the matrix, they are challenging to adapt for cell-matrix constructs, especially for the incubation times needed to observe cell-mediated matrix remodeling (e.g., matrix compaction, fibril reorganization, collagen synthesis; refs. 2
, 9
, 10
, 19
20
21
). At the macroscale, these challenges have been overcome using cell-matrix constructs cultured in tubular geometries and mechanically characterized under defined mechanical and biological conditions (19
, 22
, 23)
. The resource costs associated with these experiments have so far largely precluded exploration of the effects of soluble mediators, small molecule inhibitors, and genetic modifications outside the known fibro-proliferative molecular pathways.
A complementary approach that could reduce resource costs and enhance discovery efforts is to use micron-scale particles embedded within the matrix as mechanical probes. The approach is similar to microrheometric methods for tracking spontaneous (24
25
26
27)
or forced (28
29
30
31)
motions of micron-scale particles within or on the surface of polymer, protein, and cellular preparations. Velegol and Lanni (29)
first explored the tracking of forced bead oscillations as a method to mechanically characterize collagen gels. Though successful, the method relied on laser trapping and oscillating individual beads dispersed throughout a gel and calculating local mechanical properties, a cumbersome task given the regional heterogeneity of bead behavior within the collagen gels. To overcome this limitation, we adapted a previously developed method (28
, 32)
to simultaneously measure the collective behavior of a population of beads embedded within a type I collagen matrix; the method provides a single robust readout of matrix mechanics suitable for screening responses to a variety of interventions. Here we validate this approach in collagen gels, extend it to matrix constructs undergoing cell-mediated remodeling, and use the assay to screen the effects of selected cytokines on cell-mediated changes in matrix mechanical properties. We also measure changes in expression of four matrix remodeling-associated genes under identical treatment conditions for comparison with changes in matrix mechanical properties.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Cell culture
LL-24 adult human fibroblasts were grown in a monolayer culture at 37°C and 5% CO2 in F12K plus 15% fetal bovine serum (FBS), as recommended by ATCC. The medium was supplemented with 100 U/ml penicillin and 100 µg/ml streptomycin. Cells between passages 9 and 11 were used for all experiments. Adult lung fibroblasts were used in this study based on their involvement in pulmonary fibrosis and other matrix disorders of the adult lung (3
, 33)
.
Collagen gel preparations
Vitrogen 100 was mixed with 10x phosphate-buffered saline (PBS) and 0.1 M NaOH in a ratio of 8:1:1 to prepare isotonic solutions at pH of 7.4 ± 0.2 and a final collagen concentration of 2.0 mg/ml. Ferrimagnetic beads (4.5±0.45 µm diameter) in several batches were produced and characterized at the Harvard School of Public Health. Unless otherwise indicated, beads were used at a final mass concentration of 12.5 µg per 100 µl of final collagen solution. Beads were equilibrated in 1x PBS before addition to collagen mixtures. The method used here for embedding beads within collagen gels is equivalent to a previously described method for passively adsorbing a collagen coating onto ferrimagnetic beads (34)
. The protocol described above thus results in passive adsorption of collagen onto the surface of the beads during fibrillogenesis and integration of the beads into the matrix of the collagen gel.
To create cell-matrix constructs, the neutralized dilute collagen and ferrimagnetic beads were mixed with LL-24 fibroblasts suspended in F12K medium supplemented with 5% FBS to achieve the desired final cell concentration. To prepare constructs that maintain their radial dimension during extended culture periods, 200 µl of the combined solution was aliquoted into porous polyethylene tubes punched and hollowed from porous polyethylene sheets (average pore size of 80–120 µm, dimensions of 10 mm external diameter, 6 mm internal diameter, and length of 6 mm). The bottoms and outer sides of the polyethylene cylinders were covered in Parafilm to prevent collagen leakage during fibrillogenesis and each cylinder was placed in a well of a 24-well plate. The constructs were incubated for 45 min at 37°C in room air to facilitate fibrillogenesis. When fibrillogenesis was complete, the Parafilm was removed from each cylinder and 500 µl of F12K plus 5% FBS was added to each well; this serum concentration was chosen to maximize cellular responsiveness to subsequent cytokine stimulation (35)
.
Constructs were stimulated with TGF-ß1, PDGF-BB, Et-1, IL-1ß, IL-4, and IL-13 at concentrations of 2, 20, and 50 ng/ml, covering the typical range of biological activity for these cytokines (35
36
37
38
39
40
41
42
43
44
45
46
47)
. Cytokine stimulation was initiated when the cells were fed 1 day after fibrillogenesis. Cells were then fed every 2 days with F12K plus 5% FBS and the indicated cytokines, alone or in combination. Time course experiments were continued for up to 14 days, which allowed sufficient time for slow-developing cellular processes such as matrix neosynthesis (2
, 9
, 10
, 19
20
21)
to occur and their effects on matrix mechanics to be included in the analysis.
Microrheometry of collagen gels
Magnetic bead twisting analysis of collagen gels was performed using a described device (28
, 32
, 48)
. The device exposes matrix constructs to a brief magnetic pulse, which uniformly magnetizes the beads embedded within the matrix (Fig. 1
A). The device then exposes constructs to either a constant twisting field (10, 20, or 40 Gauss) for 60 s or a sinusoidally varying twisting field applied at one of two frequencies (0.05 and 0.2 Hz) for 10 cycles. The twisting field reorients the beads, and the ensemble average rotation of the beads is calculated by measuring the change in the remanent magnetic field of the beads in the direction orthogonal to the twisting field. The remanent field strength can decay slowly due to random bead motions, so this slow decay is accounted for by measuring the decrease in magnetic field strength in time-matched control experiments. Changes in the corrected remanent field strength are the sole result of bead rotation in the twisting field (Fig. 1B
).
|
The extent of bead rotation in response to a twisting field reflects the properties of the material in which the beads are embedded; the stiffer the material, the smaller the rotation. For constant twist experiments, the material is characterized by an "apparent stiffness," which is computed from the applied torque and the resulting ensemble angular rotation of the beads. The applied torque is a product of the magnitude of the twisting field applied, and the magnetic content of the beads determined from bead rotations in a fluid of known viscosity (32
, 49)
. The measurement of cumulative bead rotation after 60 s in a constant twisting field is analogous to a shear-creep test. The apparent stiffness calculated in this manner is closely related to the material shear modulus (G) for beads that are uniformly and fully embedded within a linear isotropic material (50)
.
For oscillatory twist experiments, the twisting field Hz(t) applied to the beads is a sinusoidal function with amplitude Ha and frequency f: Hz(t) = Ha sin (2
ft). The time-varying angle of bead rotation 
(t) produced by the twisting field is computed by comparing the measured bead field B(t) to a reference (before twist) field B0: (t) = cos–1 [B(t)/B0]. The specific torque present at the bead surface T(t) is then: T(t) = c Hz(t) cos [(t)], where c is a bead calibration constant incorporating the magnetic content of the beads (49)
. The complex modulus G for the material at any frequency is defined as: G(f) = T(f)/(f). This can be represented as G(f) = G’(f) + i G"(f), where i is sqrt(-1), G’ is the storage modulus, and G" is the loss modulus (48)
.
Macrorheometry of collagen gels
The bulk elastic modulus of collagen gels was measured using a rheometer (CVOR, Bohlin Instruments, East Brunswick, NJ, USA) with a 4°, 40 mm cone and plate geometry. The rheometer is equipped with a heating unit, and samples were maintained at 37°C during testing. The elastic modulus, G’(
), and loss modulus, G"(), were measured as a function of frequency over the range of = 0.15–5 Hz in a stress-controlled mode. The maximum applied stress was 0.1 Pa. This led to a predominantly in-phase (elastic) strain with a maximum of 
0.09%. We verified that the frequency-dependent modulus did not depend on the magnitude of the applied stress until it was increased to larger than 1 Pa (with a corresponding maximum strain of 1%), and thus our measurements were in the linear regime.
Fluorescent staining of fibroblasts
Constructs were fixed with 3% paraformaldehyde, permeabilized with 1% triton X-100, and stained with Alexa Fluor 488 phalloidin and DAPI (Molecular Probes, Carlsbad, CA, USA). Filamentous actin cytoskeletal structures and nuclei were imaged using an inverted fluorescence microscope (Nikon) connected to a cooled CCD camera (Roper Scientific, Duluth, GA, USA).
Reverse transcription and real-time polymerase chain reaction (RT-PCR)
Cell-matrix constructs were freed from porous polyethylene rings and minced with forceps, then digested in collagenase solution (3 mg/ml, Invitrogen) at 37°C for 10 min. The crude digests from 8–10 constructs per condition were pooled, the fibroblasts collected by centrifugation, and total RNA was purified with a commercial kit (Rneasy, Qiagen, Valencia, CA, USA). Equal amounts of RNA (1 µg) were reverse transcribed with Ready-to-Go RT-PCR beads (Amersham, Arlington Heights, IL, USA). PCR primers (Invitrogen) targeting alpha-smooth muscle actin (
-SMA), collagen I(I), collagen I(III), connective tissue growth factor (CTGF), and glyceraldehyde phosphate dehydrogenase (GAPDH) were designed using Primer Express software (Applied Biosystems, Foster City, CA, USA) with similar melting point temperatures, primer lengths, and amplicon lengths in order to obtain similar PCR efficiency. Real-time PCR reactions used SYBR Green Master Mix (Bio-Rad, Hercules, CA, USA) to detect and quantify PCR products. The fold change in the expression of each gene in response to each treatment was calculated relative to untreated controls using the delta-delta Ct method with normalization to GAPDH (51)
.
Statistical analysis
Results from step twist experiments are presented as percent changes in apparent stiffness normalized to internal controls for each experiment. Results from cytokine stimulation dose- and time-response experiments were analyzed by Student’s t test, comparing the cytokine-induced change in apparent stiffness at each time and dose to untreated time-matched controls. The modulation of TGF-ß1 effects by other cytokines was analyzed by 1-way ANOVA with a Bonferroni correction applied for multiple comparisons. P values <0.05 were considered significant.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, the remanent magnetic field associated with the beads, measured transverse to the twisting field, declines as the beads rotate toward alignment with the applied twisting field. Upon removal of the twisting field, the detected remanent field recovers to the level of no-twist controls (Fig. 1C
), indicating that the beads rotate quickly back to their starting orientation. If the beads were not coupled to the matrix, some residual bead rotation would remain. These results demonstrate strong coupling of the beads to the collagen gel and a no-slip interaction at the bead-matrix interface.
Using the step twist microrheometry method, the apparent stiffness measured in a 20G twisting field averaged 38.7 ± 5.3 Pa in collagen gels ranging in volume from 40 to 100 µl (2 to 5 mm in thickness), with bead concentrations ranging from 20 to 100 mg/well (data not shown). Despite the relatively high mass concentration of beads in the constructs, the volume fraction of beads in the constructs was low (0.005%). The apparent stiffness of the collagen gels measured here is within the range reported for the shear storage modulus of similar concentration collagen preparations (52)
. Moreover, the robustness of apparent stiffness measures reported above, which were obtained from gels over a range of different gel volumes and bead densities, alleviated concerns that the assay might be influenced by mechanical interactions between particles or between particles and the rigid walls of the wells.
Because biological materials frequently exhibit rate-dependent (viscoelastic) mechanical behavior, we also characterized the viscoelastic properties of the collagen gels using the same microrheometry system. Collagen preparations with embedded beads were exposed to an oscillatory twisting field at one of two frequencies (0.05 and 0.2 Hz) for 10 cycles, and the resulting complex modulus was decomposed into storage (G’) and loss (G") moduli (48)
. While the measured loss modulus exhibited greater variability than did the storage modulus, both were independent of frequency within the limited range measured (Fig. 1D
). The storage modulus averaged 121 ± 18 Pa, substantially higher than the apparent stiffness measured in static twist experiments. This difference likely reflects the fact that a finite viscous loss modulus (G") corresponds to a shear modulus that decreases slowly in time in the static twist measurement, and thus produces less apparent stiffness. However, when we exposed the same collagen preparations to static and oscillatory twist measurement techniques, the difference in measures of stiffness was much smaller, indicating that variability in collagen preparations also plays a role (data not shown).
To validate the microscale magnetic bead twisting approach, we compared it to a traditional macrorheometric method. Similar collagen preparations (1.5 mg/ml) were cast in a cone and plate rheometer and exposed to oscillatory shear in a stress-controlled mode over a wide frequency spectrum (53)
. Due to differences in the two different mechanical characterization methods, we were only able to directly compare the mechanical properties at a single frequency. Nevertheless, the magnitudes of the storage and loss moduli and their relative ratios were similar in the bead twisting and macrorheometry approaches (Fig. 1D
), and both yielded results within the range of reported values for collagen preparations of similar concentration (52
, 54)
. The highly sensitive macrorheometric technique was able to detect the subtle frequency-dependent increase in G’ and G" (Fig. 1D
), but as with the bead twisting approach, the macrorheometric technique appeared to display increasing variability in G" measurements with declining frequencies.
Measuring cell-mediated changes in matrix mechanics
To study fibroblast-collagen constructs in a relatively constant geometry for extended culture periods, we carried out fibrillogenesis inside short hollow tubes machined from porous polyethylene (Fig. 2
A). The intercalation of fibrils with the porous microstructure of the polyethylene prevented radial compaction of the constructs during the 2 wk of culture we studied. The constructs were left free to compact in the axial directions; this geometry facilitated microscopic visualization of the cells (and beads) during and after extended culture (Fig. 2B
).
|
Over 7 days in culture, the cell-matrix constructs underwent cell-dependent changes in matrix storage modulus, as measured by the microscale bead twisting assay in oscillatory mode (0.2 Hz frequency). In the absence of cells, the measured stiffness of the matrices underwent a gradual decline over 7 days (Fig. 2C
). In the presence of human lung fibroblasts, the stiffness measured 1 day after seeding varied inversely with cell density. Over time, the cells increased the gel stiffness relative to no-cell controls, and attenuated or completely reversed the cell-free decline in stiffness (Fig. 2C
). The strongest response was observed at the highest cell density (100,000 cells/ml). Although this cell density is at the low end of those used in other matrix remodeling assays (13
, 20
21
22
23
, 38
, 55)
, higher cell densities sometimes led to detachment of gels from the porous polyethylene supports. Using a rough approximation for average cell volume of 1000 µm3, the starting volume fraction of cells in the constructs was 0.01% at the highest cell density used, highlighting the sparseness of the cells in the constructs. Keeping cell (and bead) densities low minimized potential bead-cell interactions and ensured that the mechanical measurements reflected matrix, not cellular, mechanical properties.
Cytokine-induced changes in matrix mechanics
We used the assay in step twisting mode to screen the effects of cytokines and growth factors on cell-mediated changes in matrix stiffness. While collagen gel mechanical properties were consistent within each preparation, they exhibited variability between preparations. Therefore, all experiments to measure cell-mediated changes in matrix stiffness were performed with suitable internal controls (cell-matrix constructs without exogenous soluble mediators) to normalize changes in matrix stiffness across experiments and conditions. We compared the effect of six different cytokines and growth factors implicated in fibrotic processes in the lung: TGF-ß1, PDGF-BB, Et-1, IL-1ß, IL-4, and IL-13 (3
, 33
, 36
, 46
, 47
, 56
57
58
59)
. Each was administered to cell-matrix constructs 1 and 3 days after fibrillogenesis, and the apparent stiffness of cell-matrix constructs was measured at 4 days (Fig. 3
A). TGF-ß1, the most widely studied profibrotic mediator, increased cell-matrix construct stiffness in a concentration-dependent fashion, reaching statistical significance at 50 ng/ml (P=0.048). IL-1ß appeared to be the second most potent factor tested, but its effectiveness peaked at the intermediate concentration of 20 ng/ml; this effect did not reach statistical significance in this preliminary experiment. Contrary to expectations, PDGF-BB did not significantly enhance the stiffness of the cell-matrix constructs at any of the three concentrations. Et-1, IL-4, and IL-13 also exhibited statistically negligible effects on matrix stiffness at all three concentrations tested.
|
To confirm and extend these results, we measured time-dependent changes in cell-matrix construct stiffness at 2, 4, and 8 days for all cytokines and at 14 days for a subset, using the 20 ng/ml concentration for each soluble factor (Fig. 3B
). This more extensive time course study, which used separate collagen preparations for each measurement group, produced consistent trends across and within treatments. As in the previous experiment, TGF-ß1 and IL-1ß exhibited the most potent enhancing effects, with statistically significant positive contributions to matrix stiffness for TGF-ß1 at days 2 (P=0.043), 8 (P=0.004), and 14 (P=0.006), and IL-1ß at days 4 (P=0.020) and 8 (P=0.017). Whereas PDGF-BB increased matrix stiffness at day 2 (P=0.0005), it decreased stiffness at days 4 (P=0.011), 8 (P=0.002), and 14 (P=0.035). The remaining cytokines exhibited more modest and variable effects on matrix stiffness, with statistically significant increases in matrix stiffness for Et-1 at day 2 (P=0.0007), IL-4 at day 8 (P=0.026), and IL-13 at day 2 (P=0.012).
Given the strong effect of TGF-ß1 on cell-matrix construct stiffness and the frequency with which inflammatory and fibrotic processes develop together, we wondered whether any of the mediators, which by themselves had only modest effects, would augment or modulate TGF-ß1’s stiffening effects. At 8 days, Et-1 and IL-13 did not significantly modify the stiffening response of cell-matrix constructs to TGF-ß1 (Fig. 4
A). In contrast, IL-1ß did augment the TGF-ß1-induced stiffening response (P=0.029, Fig. 4A
). In separate experiments, the combination of TGF-ß1 and IL-1ß exhibited cooperative to synergistic effects on matrix stiffness at days 8 and 14 (Fig. 4B
).
|
To begin the process of delineating how IL-1ß and TGF-ß1 could produce an augmented stiffening response, we tested for cooperative effects of these cytokines on the expression of four signature genes implicated in matrix remodeling processes: -SMA, collagen types I and III, and CTGF. We treated cell-matrix constructs with 20 ng/ml IL-1ß, TGF-ß1, or both cytokines at 1 and 3 days after fibrillogenesis, then at 4 days collected and pooled mRNA from 8–10 constructs from each cytokine treatment group (and a time-matched control group). Treatment with TGF-ß1 had the expected stimulatory effect on expression of all four remodeling signature genes by quantitative real-time PCR (Fig. 4C
). In contrast, IL-1ß had only a modest positive effect on collagen III and CTGF expression, and a modest negative effect on -SMA and collagen I expression. None of the genes we examined exhibited augmented changes in expression when cells were simultaneously stimulated with both IL-1ß and TGF-ß1.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 60
61
62
63)
. In contrast, parallel efforts to delineate the biological mechanisms and molecular pathways by which cells modulate matrix mechanics are limited by the dearth of tools optimized for the job. The assay described here is designed to address this need and accelerate systematic exploration of the stimuli and cellular programs that orchestrate resident cell modulation of matrix mechanics. The microscale bead twisting assay is shown here to efficiently measure collagen gel mechanical properties, extracting matrix stiffness measurements comparable to a traditional engineering approach (Fig. 1D
). In contrast to traditional approaches, the microscale assay is well suited to follow cell-mediated changes in matrix mechanics over extended durations; due to the small sample volumes and short durations required for each measurement, the assay can be used to efficiently explore the effects of a variety of molecular interventions on the integrated functional outcome of matrix stiffness.
As a demonstration of the assay’s strengths, we examined the effects of six cytokines over three concentrations and four different durations, and performed follow-up analyses of the interactions between a subset of the mediators. The cytokines we examined were chosen for their known or suspected association with tissue remodeling and fibrosis (3
, 33)
. Though much is known about the biological effects of these cytokines, surprisingly little is known about their direct effects on matrix mechanical properties, a potentially crucial functional role that has been largely overlooked. Our results demonstrate early and sustained increases in matrix stiffness in response to TGF-ß1, a more transient stiffening response to IL-1ß, and a decrease in matrix stiffness in response to PDGF-BB (Fig. 3B
). While the matrix stiffening response to TGF-ß1 concurs with findings in cell-mediated strengthening of fibrin gels (17)
, the unexpected responses to PDGF-BB and IL-1ß demonstrate the utility of the assay in uncovering new links between biological signaling and cellular modulation of matrix mechanics.
The responses to PDGF-BB and IL-1ß were striking because they contradict the expected relationship between fibro-modulatory stimuli and changes in matrix stiffness. PDGF-BB is regarded as a profibrotic peptide in the lung; it induces collagen expression in cultured monolayer fibroblasts (33
, 64)
, and inhibition of PDGF signaling attenuates pulmonary fibrosis (56)
. In our experimental system, the biological effects of PDGF-BB on human lung fibroblasts produced a decrease in the stiffness of collagen gels from 4 to 14 days. The decrease in stiffness caused by PDGF-BB may relate to the findings of Shreiber et al. (38)
, who observed that PDGF-BB decreases fibroblast traction forces while increasing migratory behavior in stressed collagen gels. These behaviors are opposite from the enhancement of traction forces by TGF-ß1 reviewed by Tomasek et al. (2)
and the increase in matrix stiffness that TGF-ß1 produced in our system (Fig. 3A
).
In contrast to PDGF-BB, the consensus view of IL-1ß is more nuanced. IL-1ß applied directly to lung fibroblasts can cause either proliferation (65)
or apoptosis (36)
, but in general IL-1ß is thought to decrease collagen synthesis by fibroblasts (33)
. These direct cellular responses to IL-1ß should not be confused with systemic effects of IL-1ß on inflammation and subsequent matrix remodeling (59)
. In our experimental system, IL-1ß stimulated human lung fibroblasts to enhance matrix stiffness and augmented TGF-ß1-induced stiffening. In contrast, neither Et-1 nor IL-13 significantly enhanced the TGF-ß1 stiffening response, even though evidence exists for cooperative effects of these cytokine combinations in other scenarios (37
, 40
, 66)
. Although IL-1ß and TGF-ß1 have been shown to cooperate in stimulating production of fibronectin (42)
, prostaglandin E2 (35)
, and cyclooxygenase-2 (35
, 39)
, they have opposite or antagonistic effects on other key fibroblast behaviors, including apoptosis, myofibroblast differentiation, and collagen synthesis (2
, 36
, 47
, 58
, 67
, 68)
. Together the complex and unexpected stiffening responses to these cytokines and cytokine combinations reinforce the notion that effects on matrix mechanics cannot be predicted by extrapolating from known effects on cell signaling, proliferation, or matrix synthesis.
Our analysis of fibroblast gene expression further underscores the disconnect between biochemical and mechanical outcomes. While our analysis was limited to four matrix remodeling-associated genes, in no case did the relative change in gene expression induced by TGF-ß1, IL-1ß, or their combination parallel the relative change in matrix stiffness induced under identical conditions. The opposing effects of TGF-ß1 and IL-1ß on -SMA and collagen I expression reported here are supported by similar findings from cytokine stimulation of traditional monolayer fibroblast cultures (36
, 58
, 67
, 68)
.
The absence of correlation between matrix remodeling-associated gene expression and changes in matrix stiffness highlights the need for a comprehensive examination of the cellular, molecular, and microstructural processes by which cells and cytokines modulate matrix stiffness. A number of candidate processes could be involved, including cellular proliferation, migration and traction force generation, matrix reorganization and cross-linking, and synthesis or destruction of matrix constituents. All of these processes likely contribute to changes in matrix stiffness, with their relative contributions evolving over time and depending on the cytokine(s) involved. The experimental tools to detect these processes are available or in development, and the theoretical models for assessing their contributions to matrix stiffness are improving rapidly (8
, 9
, 18
, 53
, 54
, 60
61
62)
. A unique strength of the microrheometric approach developed here is identifying the biological stimuli, both expected and unexpected, that generate dramatic changes in matrix stiffness most suitable for these types of detailed analyses. In the case of IL-1ß and its cooperative effect with TGF-ß1, such a detailed approach seems warranted. Intriguingly, although many effects of TGF-ß1 and IL-1ß are antagonistic, both are known to stimulate fibroblast production of glycosaminoglycans (41
, 44)
which can alter the mechanical behavior of collagen gels (14
, 15)
. Further analyses using this and other methods are needed to delineate the molecular and cellular processes that underlie cytokine-induced changes in matrix stiffness. The assay described here can play a role in these mechanistic studies by providing an efficient means to test molecular and genetic strategies targeting predicted regulators of matrix stiffness.
The approach we describe provides an efficient means to probe cellular modulation of matrix stiffness, but it is not without limitations. The experimental system of fibroblasts in a collagen gel, though widely used because it retains essential elements of cell-matrix 3-dimensional interactions, omits much of the structural and molecular complexity present in situ. Moreover, the concentration of cells and collagen in our system is low relative to most biological tissues, leading to microscale stiffness measurements that are lower than most macroscale measures of biological tissue stiffness (52)
. An open question thus remains as to whether the model system used here faithfully recapitulates the in vivo conversion of cellular signaling into matrix stiffness changes. Nevertheless, the multitude of cellular processes that must be integrated and coordinated to modulate tissue mechanics in vivo supports the use of simplified in vitro systems to help identify key candidate molecular and cellular programs.
Given the profound effects that extracellular matrix mechanical properties exert on tissue functionality and cellular behavior (2
, 4
5
6
, 13
, 69)
, there is a great need to understand how cells actively modulate tissue mechanics. The assay described here provides a new approach that should facilitate systematic investigation in this critical area. When combined with imaging and theoretical and computational approaches, a better understanding of matrix mechanics and the cellular programs that modulate them will emerge. Together these approaches and the insights they provide will bolster attempts to engineer optimal replacement tissues and improve our understanding of how cellular control of tissue mechanics might be modulated to treat a variety of pathologies.
|
ACKNOWLEDGMENTS |
|---|
Received for publication October 20, 2006. Accepted for publication January 25, 2007.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||
