Open Access
Add to BibSonomyAbstract
We explore the dynamics of cancer cell filopodia of diameters around 200 nm by using super-resolution bright-field optical microscopy. The high contrast required by the super-resolution image-restoration process is from the nanometer topographic sensitivity of non-interferometric widefield optical profilometry, rather than fluorescence labeling. Because the image-acquisition rate of this bright-field system is 20 frames/min, fast cellular dynamics can be captured and then analyzed. We successfully observe the growth and activities of the filopodia of a CL1–0 lung cancer cell. In the culturing condition, we measure that the filopodia exhibit an average elongation rate of 90 nm/sec, and an average shrinkage rate of 75 nm/sec. With the treatment of epidermal growth factor, the elongation and shrinkage rates increase to 110 nm/sec and 100 nm/sec respectively. We also find that the treatment of epidermal growth factor raises the number of filopodia by nearly a factor of 2, which implies enhancement of cell motility.
©2007 Optical Society of America
1. Introduction
Cell migration plays a crucial role in the metastasis of cancers. To investigate the factors correlated with the motility or invasiveness of cancer cells, a prerequisite work is to tell the cells with higher motility from the others. Cell morphology is the most commonly used characteristic to identify high-motility cancer cells. For example, filopodia, needle-like projections protruding from the cell edge, are thought to be closely related to cancer cell motility and invasiveness [1, 2]. However, because the diameters of ordinary filopodia are around 100–200 nm [3], precise quantifications of filopodium dynamics are difficult by using bright-field or phase-contrast optical microscopy.
Because filopodia are mainly composed of actin-filament bundles, tagging the actin molecules with fluorescent proteins helps us to identify the presence and dynamics of living cell filopodia [4]. Based on the high contrast of fluorescence images, one may employ image restoration algorithms to improve the lateral resolution to better than the diffraction limit. For example, Carrington et al. have demonstrated “super-resolution” fluorescence imaging with lateral resolution down to the 100-nm scale [5]. Restored fluorescence images are also used for verifying the transporting functions of cellular filopodia [6]. Although super-resolution fluorescence microscopy is capable of observing filopodium dynamics, intense illumination is usually required for high-speed fluorescence imaging owing to the limited excitation efficiency of fluorophores. In such a condition, phototoxic effects could reduce cell viability or lead to cell death [7, 8]. Furthermore, photobleaching of the fluorophores also makes the quantifications of filopodium dynamics difficult. In order to avoid these drawbacks of fluorescence microscopy, one has to employ other contrast mechanisms with high signal-to-noise ratios to identify the thin filopodia of living cells.
In this work we use super-resolution bright-field optical microscopy to observe the filopodium dynamics of unlabeled lung cancer cells. The high-contrast signal required by super-resolution image restoration is from the topographic variations of the specimen, rather than fluorescent materials. We employ an optical technique with nanometer depth sensitivity, non-interferometric widefield optical profilometry (NIWOP) [9], to obtain the topographic contrast. NIWOP detects nanometer-scale topographic variation along the optical axis while the sample surface is placed in the sharp linear region of the axial response curve of widefield optical sectioning microscopy [10]. On solid-state samples, the super-resolution capability of NIWOP has been demonstrated to be about one seventh of the working wavelength [11]. Here we apply the super-resolution NIWOP technique on living cells in an aqueous environment. At an image-acquisition rate of 20 frames/min, the lateral resolution of the restored images is about 120 nm. With this super-resolution bright-field technique we compare the numbers and dynamics of filopodia before and after the treatment of epidermal growth factor.
2. Method and material
2.1 Setup and operation of the super-resolution NIWOP
Figure 1 shows the setup of our NIWOP system. This setup uses a conventional optical microscope with an epi-illumination port (Nikon Eclipse L150) as the mainframe. We employ a power-regulated mercury–xenon lamp as the light source, of which the power fluctuations are smaller than ±0.1%. In order to avoid cell damage caused by the ultraviolet emission from the light source, we install bandpass filters to set the illumination spectral range within 450–550 nm. A 12-cycle/mm grid pattern is projected onto the sample as the synthesized aperture to produce optical sectioning [10]. The spatial frequency of this grid pattern projected onto the cell surface is 0.88 μm−1. Light reflected from the membrane surface is collected by a water-immersion objective (Nikon CFI Plan Apochromat VC 60× WI) with a 1.2 numerical aperture. The lateral resolution of this objective calculated by the Rayleigh criterion is 250 nm. The culture dish containing the cells is placed on a PZT-driven vertical stage (Physik Instrumente P-762.ZL), which has a 10 nm smallest step size and 0.1% linearity in closedloop operation. We use this PZT stage to place the cell into the linear region of the axial response curve of widefield sectioning microscopy. A 14-bit CCD camera with 1600 × 1200 pixels (Cooke pco.1600) is employed to capture the images. The CCD chip is cooled at −25° C. The pixel size of the NIWOP images is 125 nm, corresponding to 50% of the optical resolution provided by the objective. In order to match the over-sampling requirement of resolution enhancement [5], we re-sample the images with the bilinear interpolation method [12] before the restoration process. The pixel numbers are quadrupled such that the final pixel size is 31 nm.
To obtain a NIWOP topography, three images I 1, I 2, I 3 with the grid pattern at spatial phases 0, 2π/3, 4π/3 are captured. An optically sectioned image Isec is then produced after the grid pattern is removed with the homodyne detection principle, Isec=[(I 1−I 2)2 +(I 1 −I 3)2 +(I 2 − I 3)2]1/2[10].In this work the exposure time of each single frame is 0.6 sec. At present the capture and processing time for a raw NIWOP image is within 3 sec, which is mainly limited by the mechanical stability of the moving grid pattern. Because the slope of the axial response curve is very sharp, as we place the sample surface slightly away from the focal plane (into the linear region), nanometer topographic sensitivity is easy to achieve. The height of the topography is calibrated from the slope in the linear range of the axial response curve. The depth resolution of this system is 16 nm. More details about the NIWOP technique can be found in our previous publications [9, 13].
In Ref. [11], we performed the super-resolution image restoration by a homemade program using an expectation-maximization maximum-likelihood estimation algorithm. The iteration process of such an algorithm has been verified to shrink the point spread function [14] or to extend the width of the optical transfer function [15]. This is known as the super-resolution imaging capability. With the nanometer height sensitivity of NIWOP as the contrast mechanism, we demonstrated that the lateral resolution could be improved from 230 nm to 70 nm using a metal stripe as the specimen. The drawback of our homemade iteration program is that the computation speed had not been optimized. Because consecutive highresolution image frames are necessary for living cell observations, the image processing time must be shortened. In this work we employ a commercial software product (Scientific Volume Imaging B.V. Huygens Essential) with a similar maximum-likelihood estimation algorithm to perform the super-resolution image restoration. For the images presented in this article, we set the quality factor in the software as 0.001 as the stopping criterion of iterations.
2.2 Cell culture and sample preparation
We culture the CL1–0 lung cancer cells in Dulbecco’s Modified Eagle’s Medium (DMEM) supplemented with 0.2% fetal bovine serum and 1% antibiotic pen-strep-ampho. The cells are placed in a 35-mm plastic dish and cultured at 37° C in a 5% CO2 atmosphere incubator. No treatment to the cells is required before the NIWOP observation. During the observation the culture medium containing the cells is kept at 34° C. Higher temperature leads to larger mechanical instability caused by the thermal expansion of the stage holding the culture dish.
3. Results and discussion
3.1 Contrast and resolution improvement of the NIWOP cell images
Figure 2 shows a bright-field reflection image of a lung cancer cell. The cell edge is darker than the bottom of the dish, indicating that the reflectivity of the edge membrane is smaller than that of the dish. Nevertheless, owing to the higher index contrast of the filopodia resulting from the actin bundles inside, some large filopodia are brighter than the cell edge and the dish bottom. As we demonstrated with differential confocal microscopy, extra index contrast is actually beneficial to the improvement of the final resolution of the restored super-resolution image [16]. Because obtaining accurate membrane topography is not the purpose of this work, we do not perform the inhomogeneous-reflectivity correction processing as in the previous membrane profiling study [13]. All the super-resolution images are restored directly from the raw NIWOP data.
Fig. 2. Bright-field reflection image of a CL1–0 lung cancer cell. Zoom-in images of the region enclosed by a dashed square are shown in Fig. 3.
Figure 3 shows the zoom-in images of the region enclosed by a dashed square in Fig. 2. Filopodia are invisible in the bright-field image [Fig. 3(a)] due to the lack of contrast. Nevertheless, in the NIWOP topography we can identify 7 filopodia easily. Because the diameter of a typical filopodium is smaller than 200 nm, the optical resolution in Fig. 3(b) is barely sufficient for characterizing the dynamics of a single filopodium. Therefore we apply the image-restoration process to improve the lateral resolution. For comparison, the same restoration process is performed for both the bright-field image and the NIWOP topography. The results are shown in Fig. 4. Super-resolution image restoration cannot improve the visibility of filopodia in the bright-field image because of the low contrast. But with the topographic sensitivity of NIWOP, all the filopodia are much clearer after the restoration. We note that the background noise is also enhanced after the restoration; however, it does not affect the visibility of the filopodia in Fig. 4(b).
As a verification of the improvement in lateral resolution, we extract a smaller region with dense filopodia and compare the edge response of one filopodium before and after the restoration. Figure 5(b) shows that the filopodia in the restored image are clearly resolved. We use the 10–90% intensity edge response as a measurement of the lateral resolution. After the restoration the edge response decreases to 120 nm. Compared with the original optical resolution of 250 nm, the lateral resolution is improved by a factor of 2. Because the contrast of each filopodium is different, we would not claim that the overall lateral resolution is of the same magnitude after the restoration process. However, as long as the filopodia are separated by a distance larger than 150 nm, the super-resolution NIWOP technique provides correct results for the count of filopodia.
Fig. 3. (a). Bright-field reflection image of the edge of a CL1–0 lung cancer cell. (b) NIWOP topography. NIWOP greatly improves the contrast of filopodia (indicated by the arrows).
Fig. 4. Images after the super-resolution image restoration. (a) Bright-field reflection image. (b) NIWOP topography.
Fig. 5. (a). Raw NIWOP image of a region with dense filopodia. (b) Restored NIWOP image showing 5 filopodia. (c) The line profiles across the filopodium indicated by the dashed lines in panels (a) and (b). The 10–90% intensity edge response of the restored image is about 120 nm.
3.2 Dynamics of filopodia affected by epidermal growth factor
We compare the numbers of filopodia around the cell edge before and after the treatment of epidermal growth factor (EGF). EGF promotes the proliferation and migration of epithelial cells [3, 17]. A recent study elucidates that filopodia also play an important role for the retrograde signal transduction related to EGF [6]. Therefore we postulate that the filopodia of CL1–0 cells would exhibit abundant activity as a response to EGF. Figure 6 shows the super-resolution NIWOP image of the cell in Fig. 2 before and after EGF treatment. The EGF is obtained from Sigma Chemical Co. A contrast threshold of 20% and a length threshold of 1 μm are set as the criterion for one filopodium in the super-resolution image. In Fig. 6(a), 17 filopodia are discernible from the background. After EGF treatment for 10 minutes the number of filopodia increases to 32. This observation directly verifies the enhancement of filopodium growth caused by EGF.
The increase of filopodia implies higher activities of the actin filaments near the cell edge, and therefore the dynamics of individual filopodia could also be encouraged. We select some filopodia of the same cell and compare their elongation and shrinkage rates before and after the treatment of EGF. Figure 7 shows time-lapse images of two regions on the cell edge, each with one filopodium. Because the emergence and disappearance of the filopodia are random, we cannot compare the dynamics using a single filopodium. Hence we calculate the average elongation and shrinkage rates of 9 filopodia of this cell. Before the EGF treatment, the average elongation rate is Re = 90 ± 11 nm/sec; while the average shrinkage rate is Rs = 75 ± 6 nm/sec. After the EGF treatment, the average elongation and shrinkage rates become 110 ± 12 nm/sec and 100 ± 15 nm/sec respectively. The Re and Rs increase by about 20% and 30% as a response to EGF, indicating more vigorous actin activities in the cell. Within such short an observation period, the cell does not move laterally. This is also advantageous for accurate dynamic characterizations of filopodia.
Fig. 6. NIWOP images of a CL1–0 cell after the super-resolution image restoration. (a) Before the treatment of EGF. (b) 10 minutes after the treatment of 50 ng/ml EGF. The white arrows indicate the countable filopodia.
Fig. 7. Time-lapse images of two filopodia of a CL1–0 cell before and after the treatment of EGF. Region (a) shows the elongation and shrinkage of a filopodium before the treatment. Region (b) is the case after the treatment. The numbers under each panel represent the image-capture time in min:sec. The filopodia also vibrate as they elongate and shrink.
4. Conclusion
In this work we demonstrate that the application of super-resolution image restoration to the NIWOP technique can greatly improve the lateral resolution of bright-field microscopic images. We employ this new imaging technique to observe cellular filopodia without fluorescence labeling. Because the contrast is from topographic variation, the image-acquisition rate can be as fast as 20 frames/min, which is more suitable for dynamic observation than conventional fluorescence microscopy. We successfully achieve lateral resolution of about 120 nm on a living lung cancer cell in the culture medium. The high contrast and resolution of the restored NIWOP images make the counting of filopodia more accurate. With the treatment of EGF, we find that the number of filopodia increases by nearly a factor of 2. We also measure the elongation and shrinkage rates of filopodia and verify that EGF also accelerates the dynamics of filopodia.
The non-fluorescence observation technique developed in this work is very attractive for living cell imaging. Without the high-intensity illumination required by fluorescence imaging, cells stay in their natural states before and after the observation. The results are thus more directly related to their behavior in native tissues. In addition, the preparation cost of the specimen is minimized compared to that of the transfection of fluorescent proteins into living cells. For long-term and large-scale quantifications of cellular dynamics, this super-resolution bright-field imaging technique would be especially favorable.
Acknowledgments
We thank the National Science Council of Taiwan for the financial support of this research project (contract NSC 95-2112-M-001-047).
References and links
1. J.-Y. Shih, S.-C. Yang, T.-M. Hong, A. Yuan, J. J. W. Chen, C.-J. Yu, Y.-L. Chang, Y.-C. Lee, K. Peck, C.- W. Wu, and P.-C. Yang, “Collapsin response mediator protein-1 and the invasion and metastasis of cancer cells,” J. Natl. Cancer Inst. 93,1392–1400 (2001). [CrossRef] [PubMed]
2. O. Kovbasnjuk, R. Mourtazina, B. Baibakov, T. Wang, C. Elowsky, M. A. Choti, A. Kane, and M. Donowitz, “The glycosphingolipid globotriaosylceramide in the metastatic transformation of colon cancer,” Proc. Natl. Acad. Sci. U.S.A. 102,19087–19092 (2005). [CrossRef] [PubMed]
3. L. J. Kleinsmith and V. M. Kish, Principles of Cell and Molecular Biology (HarperCollins College Publishers, New York, 1995).
4. J. V. Small, K. Rottner, P. Hahne, and K. I. Anderson, “Visualising the actin cytoskeleton,” Microsc. Res. Tech. 47,3–17 (1999). [CrossRef] [PubMed]
5. W. A. Carrington, R. M. Lynch, E. D. W. Moore, G. Isenberg, K. E. Fogarty, and F. S. Fay, “Super resolution three-dimensional images of fluorescence in cells with minimal light exposure,” Science 268,1483–1486 (1995). [CrossRef] [PubMed]
6. D. S. Lidke, K. A. Lidke, B. Rieger, T. M. Jovin, and D. J. Arndt-Jovin, “Reaching out for signals: filopodia sense EGF and respond by directed retrograde transport of activated receptors,” J. Cell Biol. 170,619–626 (2005). [CrossRef] [PubMed]
7. M. M. Knight, S. R. Roberts, D. A. Lee, and D. L. Bader, “Live cell imaging using confocal microscopy induces intracellular calcium transients and cell death,” Am. J. Physiol. Cell Physiol. 284,C1083–C1089 (2003). [PubMed]
8. S. Landry, P. L. McGhee, R. J. Girardin, and W. J. Keeler, “Monitoring live cell viability: Comparative study of fluorescence, oblique incidence reflection and phase contrast microscopy imaging techniques,” Opt. Express 12,5754–5759 (2004). [CrossRef] [PubMed]
9. C.-H. Lee, H.-Y. Mong, and W.-C. Lin, “Noninterferometric wide-field optical profilometry with nanometer depth resolution,” Opt. Lett. 27,1773–1775 (2002). [CrossRef]
10. M. A. A. Neil, R. Juskaitis, and T. Wilson, “Method of obtaining optical sectioning by using structured light in a conventional microscope,” Opt. Lett. 22,1905–1907 (1997). [CrossRef]
11. S.-W. Huang, H.-Y. Mong, and C.-H. Lee, “Super-resolution bright-field optical microscopy based on nanometer topographic contrast,” Microsc. Res. Tech. 65,180–185 (2004). [CrossRef]
12. J. C. Russ, The Image Processing Handbook (CRC Press, Boca Raton, 1995), Chap. 3.
13. C.-C. Wang, J.-Y. Lin, and C.-H. Lee, “Membrane ripples of a living cell measured by non-interferometric widefield optical profilometry,” Opt. Express 13,10665–10672 (2005). [CrossRef] [PubMed]
14. M. Schrader, S. W. Hell, and H. T. M. van der Voort, “Potential of confocal microscopes to resolve in the 50–100 nm range,” Appl. Phys. Lett. 69,3644–3646 (1996). [CrossRef]
15. J.-A. Conchello, “Superresolution and convergence properties of the expectation-maximization algorithm for maximum-likelihood deconvolution of incoherent images,” J. Opt. Soc. Am. A 15,2609–2619 (1998). [CrossRef]
16. C.-H. Lee, H.-Y. Chiang, and H.-Y. Mong, “Sub-diffraction-limit imaging based on the topographic contrast of differential confocal microscopy,” Opt. Lett. 28,1772–1774 (2003). [CrossRef] [PubMed]
17. A. Wells, “EGF receptor,” Int. J. Biochem. Cell Biol. 31,637–643 (1999). [CrossRef] [PubMed]
