The effects of cancer progression on the viscoelasticity of ovarian cell cytoskeleton structures

Article Outline

• Abstract
• Graphical Abstract
• Methods
  • Cell cultivation and sample preparation
  • Atomic force microscopy
    • Cell elasticity
    • Eelastic parameter extraction and theory
  • Cell viscoelasticity
    • μcell parameter extraction and theory
  • Cytoskeletal confocal imaging
  • Statistical analysis
• Results
• Discussion
• Acknowledgment
• References
• Copyright

Abstract 

Alterations in the biomechanical properties and cytoskeletal organization of cancer cells in addition to genetic changes have been correlated with their aggressive phenotype. In this study, we investigated changes in the viscoelasticity of mouse ovarian surface epithelial (MOSE) cells, a mouse model for progressive ovarian cancer. We demonstrate that the elasticity of late-stage MOSE cells (0.549 ± 0.281 kPa) were significantly less than that of their early-stage counterparts (1.097 ± 0.632 kPa). Apparent cell viscosity also decreased significantly from early (144.7 ± 102.4 Pa-s) to late stage (50.74 ± 29.72 Pa-s). This indicates that ovarian cells are stiffer and more viscous when they are benign. The increase in cell deformability directly correlates with the progression of a transformed phenotype from a nontumorigenic, benign cell to a tumorigenic, malignant one. The decrease in the level of actin in the cytoskeleton and its organization is directly associated with the changes in cell biomechanical property.

From the Clinical Editor

The authors have investigated changes in the viscoelasticity of mouse ovarian surface epithelial (MOSE) cells and demonstrated that ovarian cells are stiffer and more viscous when they are benign.

Graphical Abstract 

Key words: Atomic Force Microscopy, Viscoelasticity, Biomechanics, Cytoskeleton, Ovarian cancer

According to the American Cancer Society studies, cancer is the second most common cause of death in the US, exceeded only by heart disease; it accounts for nearly 1 of every 4 deaths and claims the lives of more than 1,500 Americans every day. The National Cancer Institute estimates overall costs of cancer may reach $158 billion by the year 2020. Deciphering the molecular events that lead to cancer initiation and progression for cancer diagnostic and treatment decisions based on individual genetic fingerprints has been a priority for many researchers in the last decade. However, the multigeneic cause of cancer and the genetic heterogeneity that determines disease severity and response to treatment has been a challenge and affected the success of the investigations.

Technological advances over the past few decades in the fields of nanotechnology, engineering and medicine have encouraged the investigation of the mechanical properties of individual cells to better understand human disease as well as the cell itself. It has been shown that transformed cells are often distinguishable from normal and healthy cells in many aspects, including cell growth, adhesion, morphology and organization of the cytoskeleton structure.¹, ² Therefore, the study of cell biomechanics (elastic and viscous behaviors), which have reportedly been linked to cell structure, is crucial for the development of disease-treating drugs and detection methods.³ As we further characterize the physics of cells, distinct properties could become potentially powerful tools in cancer diagnosis as well as treatment and therapy control.

Due to their “dynamic” nature, biological cells have an intricate microarchitecture that reacts and adapts to changes in both the chemomechanical environment and the disease state.³, ⁴ To date, various groups have made use of different techniques, including magnetic beads,⁵ optical tweezers,⁶ magnetic twisting and pulling cytometry,⁷, ⁸ optical stretching rheometry, micropipette aspiration and scanning force microscopy.³ However, the atomic force microscope (AFM) that was invented by Binnig, Quate and Gerber in 1986,⁹ has especially gained popularity in the task of cell structure characterization because of its unique nanoscale precision and ability to analyze live cultured cells.¹⁰ Moreover, because contrast in AFM originates entirely from the interaction force between the cantilever tip and sample, sample staining or fixation are not required.¹¹ The AFM has thus far proved useful in identifying key differences between nontransformed and cancer cells by use of nanoindentation.

Several distinct cell lines in both human and murine species have been assessed in terms of their elastic responses to force stimulation; namely, human bladder,¹² prostate,¹³ breast,¹⁴ mesothelial,¹⁵ cartilage¹⁶ and blood¹⁷ as well as murine ovarian¹⁸ and fibroblast tissues¹⁹ (Table 1). These cell lines were used to investigate the effects of cancer on individual cell structure. All groups have suggested that cancerous cells are “softer” or deform at a higher rate than their healthier, nontransformed counterparts by means of calculating their respective elastic moduli.

Table 1. Summary of previous works on the effects of cancer on cell elasticities via force microscopy

⁎Indentations performed on cell nucleus area.

Among the various cancer types prevalent in women, ovarian cancer is said to be the fifth most lethal cancer type.²⁰ It is a disease of older women and often diagnosed late, when the disease has already progressed and metastasized, at which point neither surgical nor chemotherapeutic treatment is very effective. Due to this nature, medical sources comment on this cancer type as the silent killer.

Ovarian cancer progression is accompanied by changes in gene expression levels that also affect cellular shape and architecture. Although research in this field is advancing, to date, there are no accounts of early detection for the human ovarian cancer. Moreover, no information exists about the mechanical properties of both malignant and benign human ovarian cells and the time course of changes, due to the lack of available and workable cell lines. Cell viscoelasticity has not sufficiently been determined quantitatively in any ovarian cancer progression model.

Recently, we have developed a primary mouse cell model for progressive ovarian cancer derived from C57BL/6 mice, relying on the spontaneous transformation in mouse ovarian surface epithelial (MOSE) cells in cell culture. During passaging, the cells undergo progressive changes in their growth rate and cell size, lose contact inhibition and acquire the capacity for anchorage-independent growth and tumor formation in vivo. According to their phenotype, they were categorized into early-benign-, intermediate- and late-aggressive stages of the disease.²¹ Thus, the syngeneic MOSE model represents a valid and novel alternative to human cell lines providing the early-, intermediate- and late-aggressive stages of ovarian cancer. Changes in the expression of cytoskeleton genes and their regulators and an increasingly disorganized cytoskeleton were observed during MOSE cell progression that may be associated with the aggressive phenotype.²² The focus of this fundamental study is to determine whether the differences in the mechanical properties of MOSE cells can be used as a potential marker for cancer detection. We have made use of the AFM method in connection with mathematical models to calculate both the cell elastic and viscoelastic parameters. While performing controlled examinations on MOSE cell lines representing early, intermediate and late stages of ovarian cancer, we have attempted to correlate the changes in elasticity and viscosity to disease state and the subsequent changes in the cell cytoskeleton structure.

Back to Article Outline

Methods 

Cell cultivation and sample preparation 

The MOSE cell lines representing early (passage no. 15 – 25), intermediate (passage no. 75 – 80), and late stages (passage no.155 – 171) of ovarian cancer were generated as described in our previous studies.²¹ Before the experiments, cells were cultivated in plastic T-75 (75cm² area) culture flasks. The environment used to nurture the cells consisted of high glucose Dulbecco's modified eagle's medium (DMEM-HG), which included 40 mL/L (4%) fetal bovine serum (FBS), 3.7g/L of sodium bicarbonate (HCO3), 10 mL/L of insulin-transferin-selenium (ITS), and 10 mL/L (1%) of penicillin-streptomycin solution. For AFM experimentation, cells were grown and harvested in incubators at 37°C in humidified 5% CO₂ and later plated with the density of 1x10⁵ cells/slip on 0.15 mm thick, 12 mm² glass coverslips coated with 0.1 mg/mL collagen type IV (Sigma, St. Louis, Missouri) for 24 – 30 hours. The density of the cells was maintained consistently during all the experimentation to obtain an adequate number of individual cells for the movement range of the AFM cantilever. In addition, 40 μl/3mL of HEPES at 1M concentration was added to the samples to help maintain a physiological pH of 7.2 during the experimentation. The final pH concentration of HEPES in the culture medium was 13.5 mM and was stable for more than 2 hours, which was sufficient for all AFM tests.

Atomic force microscopy 

Cell elasticity 

Measurements were performed using the force-curve technique on a Multimode V SPM (Veeco Instruments, Santa Barbara, California) with integrated optical microscope. Olympus TR400PSA V-shaped SiNi cantilevers with approximate spring constant values of ∼ 0.02 N/m were employed in all AFM experimentations; exact spring constant values were measured via thermal tuning method. The probes were modified by attaching glass spheres (Duke Scientific, Waltham, Massachusetts) of ∼10μm diameter onto the cantilever free end with two-part epoxy (Miller Stephenson, Sylmar California), which helped the cells remain minimally damaged during contact. In addition, cell-deformation nonlinearity was reduced due to a more homogenous contact between the cells and the probe. The exact diameter of the glass sphere and its attachment location were identified using a HIROX KH-7700 3D Digital Video Microscope. All cell-elasticity measurements were performed in DMEM-HG culture medium by using a MTFML standard contact mode fluid cell at room temperature (∼24°C).

Probes were positioned at the cells' nuclei proximities under optical control, and force curves were acquired at a sample rate of 5 kHz and constant approach velocity of ∼0.5μm/s (Figure 1). At significantly higher speeds (i.e., 10μm/s), the speed-dependent hydrodynamic forces acting on the cantilever will ultimately affect the tip-sample interaction.²³ Under the <1μm/s approach velocity setting condition, the movement is slow enough that viscous contributions are small, and force measurements are dominated by elastic behavior.²⁴, ²⁵ A maximum force trigger of 1.5 ± 0.3 nN was implemented for all force curves to maintain a basis for comparison among the cell lines (Figure 2, B).

• 
  • View Large Image
  • Download to PowerPoint

• Figure 1. 

  Optical microscope view screenshot taken during actual AFM experimentation. All indentations were performed on randomly selected cells that had visually “spread” morphologies.

• 
  • View Large Image
  • Download to PowerPoint

• Figure 2. 

  (A) Explanatory graph showing the full AFM stress relaxation procedure. (B) Force curves obtained via force curve technique on MOSE cells (C) Stress relaxation response curves on MOSE cells. It was visually possible to distinguish between the cell responses because of the precise sensitivity of the AFM.

E_elastic parameter extraction and theory 

Cell elastic modulus (E_elastic), was calculated by applying the Hertz model contact theory²⁶, ²⁷ to the AFM force curves. Sneddon's modification to the model was used to describe the behavior of elastic half-space bodies being subject to increasing loads by an infinitely hard surface.²⁸ The model assumes that the surface is continuous, frictionless and incompressible at small deformations,¹⁸ which does not hold entirely true for biological cells. However, the Hertz contact model is sufficient to yield a general idea of cell elasticity trends and has been used by different research groups for characterizing cell stiffness.²⁹ AFM indentation was limited to about 500 nm for the force curves to calculate the elastic moduli. This amount generally corresponded to about 15 – 30% of the total cell heights as observed in our scanning electron microscopy results on the MOSE cell line in our previous studies.³⁰ Putting limitations on the indentation depth is an issue that is widely discussed among researchers in this field, stemming from the argument for the validity range of the Hertz model. In general, the findings of researchers have led to one opinion that the Hertz model is valid only for indentation ranges that were equivalent to about < 10% of the cell thickness, and on the other hand, reports have indicated that the same model was reasonably valid for all ranges of indentation from about 4% up to roughly 30%.^25,31, ³², ³³, ³⁴ In light of these claims, we found it feasible to extend the indentations beyond 10 – 15% of the cell thickness to calculate the biomechanical properties of the MOSE cells.

Because the tip that comes in contact with the cell sample is spherical in shape, the relationship between the force and indentation is described by the following expression:

where F is the force, R is the radius of the contacting sphere, δ is the indentation and E⁎ is the relative elasticity term:where ν is Poisson's ratio and E is the elastic modulus. Following the works of Nikkhah et al¹⁹ and Darling et al,³⁴ the final form of Hertz equation can be described by³⁰:

because the Hertz model assumes linearity, this equation can be analyzed as a line in the (F^2/3, z-d) plane, from which the E_elastic parameter may be directly extracted from the slope, and the exact contact point (z₀, d₀) from the intercept value of that expression. The semi-automated mathematical calculation of this point is vital for the accuracy of the Hertz fitting and has been proposed by Guo et al.³⁵ All the force curve analysis which includes the linear curve fitting process as well as the Elastic modulus (E_elastic) calculations were made via MATLAB 2009b software.

Cell viscoelasticity 

Viscosity is an important characteristic of a material because, in reality, all materials exhibit some form of time-dependent strain.³⁶ Biological materials, in particular, are known to be able to both store and dissipate applied mechanical energy within their structures through internal frictional interactions, which are highly dependent on the rate of deformation.¹¹ Therefore, to be complete in the structural analysis of biological cells, it is imperative that an analysis or design involving human tissue must incorporate their viscoelastic behavior.

Cell measurements for this parameter were performed by acquiring stress-relaxation responses using the force scripting technique on a Dimension 3100 AFM (Digital Instruments, Plainview, New York) with PicoForce functionality. A closed loop z-position feedback control on the AFM helped to maintain a constant cantilever displacement during experimentation (Figure 2, A). All cell viscosity measurements were performed within a DMEM-HG culture medium fluid meniscus between the AFM fluid probe holder and cell sample at room temperature (24°C).

Indentations were done at the cell nucleus proximity but with a constant approach velocity of 5μm/s, to approximate a step displacement that was appropriate for the stress- relaxation model.³⁴ The data for each stress relaxation curve were sampled at 5 kHz for a total of 60 s upon the cantilever's indentation. A maximum force trigger of 3.0 ± 0.5 nN was implemented to designate the point at which the cantilever indentation halted and the corresponding z-scanner's position was held constant to record the change in deflection of the cantilever as a function of time on the sample (Figure 2, C).

μ_cell parameter extraction and theory 

A viscoelastic behavior of a material can be determined by measuring its stress-relaxation response to a designated step displacement. Specifically, a cell's apparent viscosity (μ_cell) can be extracted by applying the Hertz contact model for viscoelastic testing of materials.¹⁶

Following the work proposed by Darling et al,¹⁶ the use of an appended transient response function will give yield to the following expression:

where E_r is the relaxed modulus, δ₀ is the total deformation of the sample before stress relaxation, and both τ_σ and τ_ɛ are the relaxation time constants for load and deformation, respectively. Rearranging the terms:

This expression is an analytical equation of the force as a function of time. By fitting Eq. 5 to the force vs. time data, one can directly extract values for E_r, τ_σ, and τ_ɛ. These parameters describe a cell's viscoelastic response as a standard linear solid in the form of a spring-dashpot system in parallel with a spring. Among the several parameters that can be derived from the time constants and relaxed modulus, the apparent viscosity (μ) of the cell sample can be approximated by the following expression:

During stress relaxation, we assumed an approximately constant strain on the cells while the force was recorded over time. This is because the change in cantilever deflection during relaxation was small enough to be considered negligible when compared to the overall indentation up to the maximum force trigger value. The stress relaxation curves analysis, which includes the viscoelastic curve-fitting process, as well as the parameter calculations, were done via MATLAB 2009b software.

Cytoskeletal confocal imaging 

Immunofluorescence staining was performed on both the MOSE late and early stages to investigate their cytoskeletal organization. Our group wanted to specifically look into actin filaments in the MOSE cell line at different stages as it has been shown that actin filaments affect cell elasticity and not microtubules primarily.¹⁹, ³⁷ For the same reason, the staining of vinculin and tubulin, which are intermediate fibers that also form part of the overall cell structure, was not emphasized in this study. Cells were grown on glass coverslips, fixed with 3% paraformaldehyde followed by permeabilization with 0.5% triton X-100 and quenching with 50 mM glycine. After incubation with the primary antibodies, the cells were incubated with AlexaFluor 488 (Molecular Probes, Eugene, Oregon) or TRITC- (Sigma) conjugated secondary antibodies and mounted with DAPI-containing ProlongGold (Molecular Probes, Eugene, Oregon). Images were taken on a Nikon TE2000 confocal microscope using the NIS Elements software.

Statistical analysis 

Every indentable cell was selected at random as long as it had a spread/healthy morphology, irrespective of cancer phenotype, and was subject to a single indentation for either single force curves or stress-relaxation tests. No single cell was indented a second time, as it has been shown that cell morphology will change on contact.³⁸ Shapiro-Wilks tests were conducted to analyze the data for normality distributions. Two sample-independent t-tests were used to see if there were significant differences between the parameter results of each cell line. A 95% confidence interval (P < 0.05) was applied to assess the degrees of difference between the results of all the cell lines.

Back to Article Outline

Results 

Several AFM experiments were conducted to best approximate the results with the least amount of variability possible. Cell elasticity measurements on the early, intermediate, and late-stage MOSE cells were performed on samples sizes of 126 – 137 cells (n = 126 – 137). Cell viscoelasticity measurements were performed on sample sizes n = 30 – 33 cells (n = 30 – 33). Both versions of the Hertz contact model theory that were fitted to the acquired respective data gave yield to high correlation coefficients (0.85 ≤ R² ≤ 0.99) to suggest that theory matched satisfactorily well with experimental data.

The cell elastic property is best described by the elastic modulus, E_elastic. The parameter results were assessed in terms of cell population distributions (Figure 3).

• 
  • View Large Image
  • Download to PowerPoint

• Figure 3. 

  MOSE cell line histograms depicting the change in distribution of recorded elastic modulus from one stage of cancer to another.

The AFM indentations on the early-stage MOSE cells resulted in a broad distribution of values for E_elastic and were best described by a log-normal distribution scheme. It was interesting to observe that some cells gave yield to values that were about an order of magnitude larger than some cells in the same experiment, suggesting that this stage of the cell line is still heterogeneous with regard to elasticity because it was possible to detect a very deformable cell in proximity to a more rigid one. Intermediate stage MOSE cells produced a similar but slightly closer distribution, convincing that more cells in this stage correlated to lower elastic moduli. Finally, the late-stage MOSE cells showed a much sharper distribution, giving notice to the large number of cells that corresponded to relatively low (< 1 kPa) E_elastic values. Overall, a noticeable shifting pattern is evident in the distribution of cells from one stage of the cell line to the next. This indicates that an increasing number of low E_elastic cells are present in a cell population when transforming from a benign to an aggressive stage of the same cell line. The peak values that derive from the fitted Gaussian distribution can also be used to notice this change.

Cell viscosity (μ_cell) was used to represent a cell's viscoelastic response to force stimulation. Indentations and stress-relaxation tests were also analyzed via frequency distributions (Figure 4).

• 
  • View Large Image
  • Download to PowerPoint

• Figure 4. 

  MOSE cell line histograms depicting the change in distribution of recorded viscosity rates from one stage of cancer to another.

The cell-viscosity rate distribution was similar to that of the elastic modulus results. It is noticeable that higher rates of viscosity are present in the early, benign stage of the MOSE cell line as opposed to an increasingly homogeneous response of cells in the low range of viscosity in the later stages. Again, the peak values that arise from each stage's Gaussian distribution show this trend. The statistical analysis suggested a log-normal distribution scheme that best represented the viscosity trends above.

The elastic modulus and the viscosity rate parameters were also investigated in terms of their average and logarithmic value over each population of cell line (Figure 5). The schematic illustrates the trends from one disease state to another.

• 
  • View Large Image
  • Download to PowerPoint

• Figure 5. 

  Depiction of change in both elastic modulus and apparent viscosity parameters for the MOSE cell line during cancer progression.

A noticeable downward shift was detected for both the apparent viscosity and elastic modulus parameter values when cells transitioned from a benign state to a more aggressive, tumorigenic one. The average elastic modulus value showed about 50% decrease during the progression of MOSE cells from ∼1.1 kPa in the early MOSE cells to 0.55 kPa in the late MOSE cells. Likewise, the average apparent viscosity value showed a 65% decrease from 147 Pa-s to about 51 Pa-s. For both biomechanical properties, we also noticed a decrease in the standard deviation values, which correlate to the increasing degree of homogeneity of the cells' response in the aggressive stages. The intermediate stage MOSE cell line best served to validate the trends for both cell properties as it seemed to be in between the values for the other stages (0.8 kPa for E_elastic and 103 Pa-s for μ_cell).

The statistical significances were assessed after transforming the data into the logarithmic scale and we concluded that significant differences were present in terms of average apparent viscosity between the early versus late stage (P < 0.0001), and the intermediate versus late stage (P < 0.0001). Similarly, in terms of average elasticity, significant differences were observed between the early versus intermediate stage (P < 0.0001), the early versus late stage (P < 0.0001), and finally the intermediate versus late stage (P < 0.0001).

A series of additional tests were devoted to the validation of the output of parameter results. A total of 30 indentations (n = 30 cells) were performed for each session. The results showed that 15% – 20% error margins were associated with each cell line's average output, encouraging the hypothesis that there was a gap or difference between cell responses dependent on their stage. For example, there was a clear range between the lowest recorded early-stage cells output (0.886 kPa) and the highest recorded late-stage cells output (0.618 kPa). The intermediate-stage cell line output also situated comfortably between these values.

Noticeable differences were seen for all viscoelastic property parameters for the MOSE cell line (Table 2). The means as well as standard deviations showed a decrease in values, which reflects the observations through the obtained histograms.

Table 2. Elastic and viscoelastic biomechanical properties of the MOSE cell line (mean ± SD) in the progression of malignancy

A set of experiments were devoted to the cytoskeleton imaging of the early- and late-stage MOSE cells (Figure 6).

• 
  • View Large Image
  • Download to PowerPoint

• Figure 6. 

  Confocal images of MOSE early- and late-stage cells. A 20x objective was used in air to obtain these results. The images show the early-stage cells as having a much denser concentration of actin stress fibers when compared with their diseased counterparts.

The confocal images confirm there is a progressive dysregulation of the actin cytoskeleton, changing both the thickness and the organization of the actin fibers, which leads to its lesser concentration. These results correlate well with the changes in the general morphology of the early- and late-stage MOSE cells and changes in their cytoskeletal organization in the works of Creekmore et al, where it was found that a 78% reduction in F-actin concentration resulted from the transformation from early stage to late stage.²²

Back to Article Outline

Discussion 

The acquired AFM experimental data shown above serve best to yield the following claims: Transformed and invasive MOSE cells are more deformable and less viscous in comparison with their benign counterparts. When generalized to large populations, the MOSE cell lines become increasingly more homogeneous in their response to the AFM indentations in the progression of disease. In effect, an increase in cell deformability directly correlates with the progression of a transformed phenotype from a nontumorigenic, benign cell to a tumorigenic, malignant one. Elastic and viscoelastic deformation characteristics are well defined and expressed using the Hertz contact theory model at indentations up to 30% of cell height. This suggests that sample substrate had no effect in the biomechanical property calculations.

Changes in the cytoskeletal organization have been associated with tumorigenesis.³⁹, ⁴⁰ We have shown previously that gene expression levels are changed during the progression of the MOSE cells with an enrichment of genes in the cytoskeleton functional category.²² Specifically, genes in the actin and actin-binding regulating category exhibited the most drastic changes. This was also reflected in altered protein levels, but, most important, was accompanied by a progressive disorganization of the cytoskeleton. Actin filaments gel the cell periphery, provide the highest resistance to deformation and are the most rigid of the three main components of the cytoskeleton components.³ Because it has been shown that the nucleus region is significantly softer and deformable than the periphery,⁴¹ this could directly connect our imaging results and the AFM mechanical measurements obtained as a result of indenting the cell nucleus regions.

Research on the nature of a cell's viscous response generated novel ideas; recent studies on the effect of cytoskeleton disrupting drugs (i.e., colchicine, paclitaxel, and nocodazole) have demonstrated that cells respond with a compensatory mechanism within their structure in the face of exposure and/or stimulation.⁴², ⁴³, ⁴⁴ Furthermore, it was revealed that cells that were subject to microtubule dissociating drugs had maintained their structural integrity and showed enhanced rigidity because of increased F-actin polymerization and subsequent increase in F-actin concentration within the cell structure in response to disrupted microtubulin. This phenomenon also resulted in a 30% increase in cell viscosity and cortical tension was increased by about 18%.⁴³ It can be concluded from these findings that increased F-actin concentration is associated with increased cell viscosity. Therefore, the higher concentration of F-actin in the early-stage MOSE cell structure is most likely to be the cause of higher rates of apparent viscosity in comparison with later stages.

Janmey et al conducted a study of the major cytoskeletal fibers' stress resistance and found that actin filaments can sustain much higher stresses and withstand deformations at a much higher rate than other fibers in the cell structure. Furthermore, actin filaments tended to result in larger storage (elastic) moduli with increasing concentration.⁴⁵ Actin filaments were also shown to have elastic moduli of about ∼2.3 GPa, as found with the method by Wagner et al, suggesting that these fibers were less elastic than rubber (E = 0.001 GPa), but more flexible than bone (E = 20 GPa).⁴⁶ Because F-actin is thought to provide the highest resistance to deformation until a certain critical strain value³ and has a very rigid composition, these fibers are thought to cause differences in rigidity between the individual cell lines. In effect, a low elastic modulus and apparent viscosity is attributed to a lower concentration of the F-actin in a cancer cell's cytoskeletal structure. Our findings confirm these associations and suggest that the cytoskeleton affects the biomechanical properties of cells. Therefore, changes in these properties can be related to the motility of cancer cells and potentially their invasive potential because it allows cancer cells to deform, squeeze and migrate through size-limiting pores of tissue or vasculature onto other parts of the body.⁴⁷, ⁴⁸

It is noteworthy that our findings are a result of the applied model and approach as wwell as the loading rate settings during the AFM experimentations. For example, our studies showed that the biomechanical response changed noticeably as an increased approach velocity, and applied force led to a general increase in all elastic moduli in a proportional manner; these findings were also seen in other groups' works.⁴⁹, ⁵⁰ Furthermore, the Hertz contact model is only one of the proposed engineering approaches, and currently no proven model exists to precisely represent cell mechanics without assumptions.

Despite this source of variability, it is beneficial to identify changes in viscoelastic parameters due to cancer progression to aid the understanding of events that cause or support tumor growth and metastasis. To date, several cell lines have been analyzed in terms of their elastic modulus properties. However, their viscoelastic behavior is yet to be fully examined and/or understood. Darling et al tested the viscoelastic behavior in three well-characterized human chondrosarcoma cell lines that each reflected a different degree of invasiveness and malignancy (JJ012, FS090, and 105KC).¹⁶ Their finding that viscoelastic behavior changes with increasing malignancy is in general agreement with our own. However, these cell lines do not represent a progressive model like the MOSE model, and established cell lines may greatly adapt their geno- and phenotype to cell culture conditions.

Much more work is needed in the field of viscoelasticity of cells. The development of effective medical devices for the detection and diagnosis of cancer is highly dependent on the degree of realistic modeling of the cell samples. Although the work of previous groups emphasized solely the elastic modulus of cells, research into the individual cell characterization has much promise in the fight of cancer.

The results we obtained indicate that the progressive disorganization and pathophysiological alterations of the cytoskeleton architecture greatly affect their biomechanical properties. This confirms earlier reports and suggests that our results are not specific to our model but could indicate a general trend in the changes of the morphology and mechanics of cells due to cancer progression. The factors that are responsible for the differences in biomechanical properties and invasiveness among the ovarian cell lines are yet to be determined. However, the findings in this study, along with those of other groups, support the idea that highly invasive cancer cells are characterized by unlimited mitotic rates, higher migration rates¹⁹,and lower elastic and viscous properties that may be associated with significant differences in cytoskeleton structure. Although further work is needed to enhance our understanding of the complex architecture of living cells and the interaction of molecular and biomechanical events, we believe that our studies will contribute to the development of novel techniques and devices for clinical utility in cancer risk assessment, cancer diagnosis and treatment efficacy.

Back to Article Outline

Acknowledgment 

Award No. ECCS-IDR 0925945. The authors would like to thank the staff of the Nanoscale Characterization and Fabrication Laboratory (NCFL) at Virginia Tech for their technical assistance and providing the necessary AFM equipment.

Back to Article Outline

References 

1. Asch BB, Kamat BR, Burstein NA. Interactions of normal, dysplastic, and malignant mammary epithelial cells with fibronectin in vivo and in vitro. J Cancer Res. 1981;41:2115–2125
   • View In Article
2. Chakraborty J, Von Stein GA. Pleomorphism of human prostatic cancer cells (DU 145) in culture—The role of cytoskeleton. Exp Mol Pathol. 1996;44:235–245
   • View In Article
   • MEDLINE
   • CrossRef
3. Suresh S. Biomechanics and biophysics of cancer cells. Acta Biomaterialia. 2007;3:413–438
   • View In Article
   • MEDLINE
   • CrossRef
4. Alberts B, Johnson A, Lewis J, Raff M, Roberts K, Walter P. Molecular biology of the cell. 4th ed.. New York: Garland Science; 2002;
   • View In Article
5. Mijailovich SM, Kojic M, Zivkovic M, Fabry B, Fredberg B. A finite element model of cell deformation during magnetic bead twisting. J Appl Physiol. 2002;93:1429–1436
   • View In Article
6. Dao M, Lim CT, Suresh S. Mechanics of the human red blood cell deformed by optical tweezers. J Mech Phys Solids. 2003;51:2259–2280
   • View In Article
7. Gardel ML, Shin JH, MacKintosh FC, Mahadevan L, Matsudaira P, Weitz DA. Elastic behavior of cross-linked and bundled actin networks. Science. 2004;304:1301–1305
   • View In Article
   • CrossRef
8. Overby DR, Matthews BD, Alsberg E, Ingber DE. Novel dynamic rheological behavior of individual focal adhesions measured within single cells using electromagnetic pulling cytometry. J Acta Biomaterialia. 2005;1:295–303
   • View In Article
9. Binnig G, Quate CF, Gerber C. Atomic force microscope. Phys Rev Letts. 1986;56:930–933
   • View In Article
10. Sokolov I. Atomic force microscopy in cancer cell research. In:  Singh Nalwa H,  Webster T editor. Cancer Nanotechnology. 2007;p. 1–17
    • View In Article
11. Kumar S, Weaver VM. Mechanics, malignancy, and metastatis: the force journey of a tumor cell. J Cancer Metastatis Rev. 2009;28:113–127
    • View In Article
12. Lekka M, Laidler P, Gil D, Lekki J, Stachura Z, Hrynkiewicsm AZ. Elasticity of normal and cancerous human bladder cells studied by scanning force microscopy. Biophysics J. 1999;28:312–316
    • View In Article
13. Faria EC, Ma N, Gazi E, Gardner P, Brown M, Clarke NW, et al. Measurement of elastic properties of prostate cancer cells using AFM. Analyst. 2008;133:1498–1500
    • View In Article
    • CrossRef
14. Cross SE, Jin Y, Rao J, Gimzewski JK. Nanomechanical analysis of cells from cancer patients. Nat Nanotechnol. 2007;2:780–783
    • View In Article
    • CrossRef
15. Cross SE, Jin YS, Tondre J, Wong R, Rao JY, Gimzewski JK. AFM-based analysis of human metastatic cancer cells. J Nanotechnol. 2008;19:1–8
    • View In Article
16. Darling EM, Zauscher S, Block JM, Guilak F. A thin-layer model for viscoelastic, stress-relaxation testing of cells using atomic force microscopy: do cell properties reflect metastatic potential. Biophysics J. 2007;92:1784–1791
    • View In Article
17. Rosenbluth MJ, Lam WA, Fletcher DA. Force microscopy of nonadherent cells: a comparison of leukemia cell deformability. Biophysics J. 2006;90:2994–3003
    • View In Article
18. Carl P, Schillers H. Elasticity measurement of living cells with an atomic force microscope: data acquisition and processing. Pflugers Arch-Eur J Physiol. 2008;457:551–559
    • View In Article
19. Nikkhah M, Strobl J, De Vita R, Agah M. The cytoskeletal organization of breast carcinoma and fibroblast cells inside three dimensional (3-D) isotropic silicon microstructures. Biomaterials. 2010;31:1–10
    • View In Article
    • CrossRef
20. Jemal AD, Siegel RM, Xu J, Ward E. Cancer Statistics, 2010. CA Cancer J Clin. 2010;60:1–24
    • View In Article
    • CrossRef
21. Roberts PC, Motillo E, Baxa AC, Heng HHQ, Doyon-Reale N, Gregoire L, et al. Sequential molecular and cellular events during neoplastic progression: a mouse syngeneic ovarian cancer model. Neoplasia. 2005;7:944–956
    • View In Article
    • MEDLINE
    • CrossRef
22. Creekmore AL, Silkworth WT, Cimini D, Jensen R, Roberts PC, Schmelz EM. Changes in gene expression and cellular organization in a model of progressive ovarian cancer. PLoSOne. 2011;6:1–16
    • View In Article
23. Janovjak H, Struckmeier J, Muller DJ. Hydrodynamic effects in fast AFM single-molecule force measurements. Eur Biophys J. 2005;34:91–96
    • View In Article
    • MEDLINE
    • CrossRef
24. Lu YB, Franze K, Steinhauser C, Kirchhoff F, Wolburg H, Guck J, et al. Viscoelastic properties of individual glial cells and neurons in the CNS. Proc Natl Acad Sci USA. 2006;103:17759–17764
    • View In Article
25. Mathur AB, Collinsworth AM, Reichert WM, Kraus WE, Truskey GA. Endothelial, cardiac muscle and skeletal muscle exhibit different viscous and elastic properties as determined by atomic force microscopy. J Biomechanics. 2001;34:1545–1553
    • View In Article
26. Timoshenko SP, Goodier JN. Theory of elasticity. 3rd ed.. New York: McGraw-Hill; 1970;
    • View In Article
27. Hertz H. On the elastic contact of elastic solids. Journal für die reine und angewandte Mathematik. 1881;92:156–171
    • View In Article
28. Sneddon IN. The relation between load and penetration in the axisymmetric Boussinesq problem for a punch of arbitrary profile. Int J Eng Sci. 1965;3:47–57
    • View In Article
29. Kuznetsova TG, Starodubtseva MN, Yegorenkov NI, Chizhik SA, Zhdanov RI. Atomic force microscopy probing of cell elasticity. J Micron. 2007;38:824–833
    • View In Article
30. Ketene AN, Schmelz EM, Roberts PC, Creekmore AL, Nikkhah M, Agah M. In: Distinct biomechanical characteristics of mouse ovarian cells as a marker for cancer detection, in 1st International Conference on Applied Bionics and Biomechanics(ICABB), Venice, Italy. 2010;
    • View In Article
31. Rico F, Roca-Cusachs P, Gavara N, Farre R, Rotger M, Navajas D. Probing mechanical properties of living cells by atomic force microscopy with blunted pyramidal cantilever tips. J Phys Rev E. 2005;72:1–10
    • View In Article
32. Sen S, Subramanian S, Discher DE. Indentation and adhesive probing of a cell membrane with AFM: theoretical model and experiments. Biophysics J. 2005;89:3203–3213
    • View In Article
33. Domke J, Radmacher M. Measuring the elastic properties of thin polymer films with the atomic force microscope. Langmuir. 1998;14:3320–3325
    • View In Article
    • CrossRef
34. Darling EM, Topel M, Zauscher S, Vail TP, Guilak F. Viscoelastic properties of human mesenchymally-derived stem cells and primary osteoblasts, chondrocytes, and adipocytes. J Biomech. 2008;41:454–464
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (1002 KB)
    • CrossRef
35. Guo SL, Akhremitchev BB. Packing density and structural heterogeneity of insulin fibrils by AFM nanoindentation. Biomacromolecules. 2006;7:1630–1636
    • View In Article
    • MEDLINE
    • CrossRef
36. Meyers MA, Chawla KK. Mechanical behavior of materials. Upper Saddle River, New Jersey: Prentice Hall; 1998;
    • View In Article
37. Rotsch C, Radmacher M. Drug-induced changes of cytoskeletal structure and mechanics in fibroblasts - an atomic force microscopy study. Biophysics J. 2000;78:520–535
    • View In Article
38. Zhu C, Bao G, Wang N. Cell mechanics: mechanical response, cell adhesion, and molecular deformation. Annu Rev Biomed Eng. 2000;2:189–226
    • View In Article
    • MEDLINE
    • CrossRef
39. Lindberg U, Karlsson R, Lassing I, Schutt CE, Hoglund AS. The microfilament system and malignancy. Semin Cancer Biol. 2008;18:2–11
    • View In Article
    • CrossRef
40. Buda A, Pignatelli M. Cytoskeletal network in colon cancer: from genes to clinical application. Int J Biochem Cell Biol. 2004;36:759–765
    • View In Article
    • MEDLINE
    • CrossRef
41. Haga H, Nagayama M, Kazushige K. Imaging mechanical properties of living cells by scanning probe microscopy. Curr Nanosci. 2007;3:97–103
    • View In Article
42. Pelling AE, Dawson DW, Carreon DM, Christiansen JJ, Shen RR, Teitell MA, et al. Distinct contributions of microtubule subtypes to cell membrane shape and stability. Nanomedicine. 2007;3:43–52
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (1159 KB)
    • CrossRef
43. Tsai MA, Waugh RE, Keng PC. Passive mechanical behavior of human neutrophils: effects of colchicine and paclitaxel. Biophysics J. 1998;74:3282–3291
    • View In Article
44. Weihs D, Mason TG, Teitell MA. Bio-microrheology: a frontier in microrheology. Biophysics J. 2006;91
    • View In Article
45. Janmey PA, Euteneuer U, Traub P, Schliwa M. Viscoelastic properties of vimentin compared with other filamentous biopolymer networks. J Cell Biol. 1991;113:155–160
    • View In Article
    • MEDLINE
    • CrossRef
46. Bray D, movements Cell. from molecules to motility. 2nd ed.. New York: Garland; 2000;
    • View In Article
47. Wyckoff JB, Jones JG, Condeelis JS, Segall JE. A critical step in metastatis: in vivo analysis of intravasation at the primary tumor. Cancer Res. 2000;60:2504–2511
    • View In Article
    • MEDLINE
48. Shah-Yukovich EM, Nelson AC. Characterization of solid tumor microvasculature: a three-dimensional analysis using the polymer casting technique. Lab Invest. 1988;58:236–244
    • View In Article
    • MEDLINE
49. Li QS, Lee GYH, Ong CN, Lim CT. AFM indentation study of breast cancer cells. Biochem Biophys Res Commun. 2008;374:609–613
    • View In Article
    • CrossRef
50. Zhao M, Srinivasan C, Burgess D, Huey B. Rate- and depth-dependent nanomechanical behavior of individual licing chinese hamster ovary cells probed by atomic force microscopy. J Mater Res. 2006;21:1906–1912
    • View In Article