Quantitative analysis of focal adhesion dynamics using photonic resonator outcoupler microscopy (PROM)

Quantitative analysis of focal adhesion dynamics using photonic resonator outcoupler microscopy... Focal adhesions are critical cell membrane components that regulate adhesion and migration and have cluster dimensions that correlate closely with adhesion engagement and migration speed. We utilized a label-free approach for dynamic, long-term, quantitative imaging of cell–surface interactions called photonic resonator outcoupler microscopy (PROM) in which membrane-associated protein aggregates outcoupled photons from the resonant evanescent field of a photonic crystal biosensor, resulting in a highly localized reduction of the reflected light intensity. By mapping the changes in the resonant reflected peak intensity from the biosensor surface, we demonstrate the ability of PROM to detect focal adhesion dimensions. Similar spatial distributions can be observed between PROM images and fluorescence-labeled images of focal adhesion areas in dental epithelial stem cells. In particular, we demonstrate that cell–surface contacts and focal adhesion formation can be imaged by two orthogonal label-free modalities in PROM simultaneously, providing a general-purpose tool for kinetic, high axial-resolution monitoring of cell interactions with basement membranes. and disassembly of a FA, the size of the FA cluster varies Introduction Focal adhesions (FAs), or cell–matrix adhesions, are and is highly correlated with the level of adhesion 13,29 large specialized proteins that are typically located at the engagement and migration speed . For example, non- interface between the cell membrane and extracellular mature focal complexes (FXs) are initially formed at the 1–24 matrix (ECM) (Fig. 1a, b) . FAs are critical for sup- leading edge of the cell (e.g., in the lamellipodia area) and porting the cell membrane structure and regulating signal are usually <0.2 µm . As the lamellipodia withdraws from transmission between the cytoskeleton (e.g., actin) and the leading edge, many FXs disassemble and release transmembrane receptors (e.g., integrins) during adhesion adhesion proteins back to the inner cell body, whereas 16–24 2 and migration . Monitoring the response of FA clus- some of the FXs grow larger (typically 1–10 µm ) and ters to drugs is one important mechanism by which the assemble into mature FA clusters by recruiting adapter 19,29 action of pharmaceutical compounds may be evaluated, proteins . Once the remaining FAs are in place, they particularly where approaches that enable characteriza- may form stationary attachment points by binding to the tion to be performed with a small number of cells are ECM, and a cell may utilize these anchors to migrate over 22,25–28 especially valuable . During the dynamic assembly the ECM by pushing and pulling the entire cellular 18,21,23 body . This insight into the dynamics of FA cluster formation and dissociation has been made possible by technical advances in the field of fluorescence and super Correspondence: Brian T. Cunningham (bcunning@illinois.edu) 1 30–36 Department of Bioengineering, University of Illinois at Urbana-Champaign, resolution microscopy . Optical modalities, including Urbana, IL 61801, USA total internal reflection fluorescence microscopy, Micro and Nanotechnology Laboratory, University of Illinois at Urbana- photoactivation localization microscopy (PALM), Champaign, Urbana, IL 61801, USA Full list of author information is available at the end of the article © The Author(s) 2018 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to theCreativeCommons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. 1234567890():,; 1234567890():,; 1234567890():,; 1234567890():,; Shift Zhuo et al. Light: Science & Applications (2018) 7:9 Page 2 of 15 ab Focal adhesion Evanescent field Extra cellular matrix Photonic crystal surface Photonic crystal surface Cell non-attached Cell attached Δλ (PWS) Cell BG ΔI (PIS) Photonic crystal Cell 0.9 biosensor 0.8 0.8 0.6 624 626 628 Objective lens 0.4 Wavelength (nm) Cylindrical 0.2 lens Dichroic 0 mirror 610 620 630 640 650 Wavelength (nm) Spectrometer Polarized LED beam splitter Mirror CCD camera Tube lens Fig. 1 Principle of the molecular-dynamics for cell attachment on a photonic crystal (PC) biosensor in photonic resonator outcoupler microscopy (PROM). Schematic representation of the molecular mechanism a before and b after a live cell attaches to the PC biosensor surface. c The principle of PROM imaging system. Inset: spectra shift before and after the cell attaches to the PC surface 17,37–39 stochastic optical reconstruction microscopy, and inter- micropillar substrates) , mechanical probing of cells 40,41 ferometric PALM, coupled with fluorescence tagging of (e.g., atomic force microscopy) , and single molecular the element(s) of FA clusters via administration of fluor- techniques (e.g., tension sensors) have allowed the escently labeled antibodies or incorporation of fluorescent quantification of molecular tension forces within FA reporter genes by transfection of cells, along with progress clusters as well as FA-mediated traction and adhesion made in single particle tracking algorithms, have allowed forces. researchers to quantify FA-associated parameters, such as Understanding the dynamics of FA formation and FA areas and sizes (x–y dimensions), FA architectures changes in FA-associated parameters is beneficial not only (x–y–z dimensions), FA turnover rates, and spatio- for understanding the fundamentals of biology but also temporal distributions of FA complexes. Additionally, for the field of biosensor diagnostics and screening for 36,43,44 developments in traction force measurements (e.g., based clinical applications . Changes in FA-associated on two-dimensional (2D) hydrogel substrates or parameters, such as FA sizes and traction forces, have Reflectance (AU) Reflectance (AU) Zhuo et al. Light: Science & Applications (2018) 7:9 Page 3 of 15 been linked to critical cellular processes, including value (PIV) from the PC before and after live cell metastasis, apoptosis, and chemotaxis, as well as pathol- attachment to acquire the peak intensity shift (PIS) at 9,29,36,45–47 ogies of cancers and other diseases . As such, each local voxel volume (Fig. 1a, b). Images of the PIS monitoring the response of FA clusters to drugs, for reveal highly localized and easily observable loci of protein example, is an important mechanism by which the action clusters that correlate with the spatial distribution and 22,25–28,36 of pharmaceutical compounds may be evaluated , size of FAs observed by fluorescence microscopy. We hypothesize that the observed reduction in reflected and high-throughput approaches that enable the char- acterization of small cell populations in real time are intensity from the PC is mainly caused by the outcoupling especially valuable for these applications. Currently of resonant standing wave photons via scattering. available techniques largely make use of fluorescence Because this imaging modality operates using an inde- tagging to mark individual FA proteins, which entails pendent sensing mechanism that obtains contrast through temporal limitations imposed by photobleaching and the formation of protein clusters at the cell–ECM inter- challenges associated with accurate quantitation and face that are capable of outcoupling light from the PC 9,11,48 long-term analysis . New tools are therefore required biosensor surface, we name this technique photonic to study the dynamic behavior of FA clusters and their resonator outcoupler microscopy (PROM) (Fig. 1c). In interaction with the ECM to characterize changes in FA our previous study, we report the first observation of a dynamics in live cells in situ. However, determining the reduced and highly localized reflected intensity in the dynamic activity of a FA cluster is challenging, especially context of nanoparticles with optical absorption at the with all of the FA proteins that are simultaneously active resonant wavelength of the PC . However, these obser- during the in situ assembly and disassembly processes in vations were made with very high contrast and highly live cells. Although a variety of approaches have been localized metal absorbers (for plasmonic nanoparticles) or utilized to investigate these processes, the detailed titanium dioxide (TiO ) nanoparticle dielectric scatterers. mechanism of FA assembly and disassembly in live cells, Although the reduced reflected resonant intensity from including the variability of the FA dimension, is poorly the high-contrast surface-attached TiO scatterers (the 9,11,48 understood . For instance, fluorescent tags are often refractive index of TiO nanoparticles (n = ~2.4) is 2 TiO2 used to mark individual FA proteins, but due to the much larger than that of the surrounding water medium temporal limitations imposed by photobleaching, accurate (n = ~1.333)) was the first observation with PROM, this quantitation and long-term analysis are exceedingly dif- study is the first to use PROM to observe scattered out- ficult to perform, whereas the cytotoxicity of fluorescent coupling from very low contrast FAs in live cell mem- tags compromises the viability of the cells under study. branes (averaged cell n = 1.35–1.38) to the surrounding cell Here we describe a label-free optical sensing approach medium (n = ~1.333). Using dental epithelial stem cells that combines two optical modalities to quantify the FA- attached to a fibronectin-coated ECM surface as a associated parameters that are critical for characterizing representative example, we demonstrate that PWS and spatiotemporal distribution patterns and the strength of PIS images of the same cells display distinct and com- FA clusters in real time. In our previous studies on cell plementary information. Although the regions with the imaging by photonic crystal enhanced microscopy greatest PWS are at the cell–surface interface in which (PCEM), we utilized an imaging modality in which the uniformly distributed regions with the greatest surface reflected resonant peak wavelength value (PWV) was engagement occur, the regions with the greatest PIS measured over the imaging field-of-view to derive images represent the formation of highly concentrated protein of the peak wavelength shift (PWS) that occur when cells clusters at the cell–surface interface that are capable of 49–52 attach to the photonic crystal (PC) surface . In these scattering photons. Therefore, we introduce PROM as a studies, we describe how the engagement of the cell quantitative, dynamic, and label-free approach to observe membrane components with the surface of the PC results the formation and evolution of FA cluster areas that are in highly localized shifts in the resonant reflected wave- otherwise challenging to observe with other available length from the biosensor , as well as the design of a imaging modalities, particularly for repeated observations modified brightfield (BF) microscope that enables visua- of the same cell population for extended time periods. lization of cell–surface attachments with a ~0.6 × 0.6 µm pixel size. The PWS image sequences clearly show the Materials and methods evolution of cell attachment through the engagement of PC biosensor the lipid bilayer cell membrane and internal cell- The PCs used in this study are subwavelength nanos- associated proteins within the ~200 nm deep evanescent tructured surfaces with a periodic modulation in the field region of the PC. In this study, we demonstrate a refractive index that acts as a narrow bandwidth resonant novel and orthogonal imaging modality within PCEM in optical reflector at one specific resonance wavelength 49,50,52–55 which we measure the resonant reflected peak intensity (Fig. 1) . The high reflection efficiency of the PC Zhuo et al. Light: Science & Applications (2018) 7:9 Page 4 of 15 ab c d s n d =d +d n g s (Medium) d g s n d =d 1 s d =d g  d =0 (Slab) d =0 (UVCP) 0 0 k z z TE TM n’ y n’ (Substrate) x e fg PIV PWV 1.02 Linear Linear 0.9 1 626 0.98 0.8 625 1.33 1.34 1.35 1.36 1.37 0.8 1.34 1.36 624 626 628 630 Refractive index (AU) Refractive index (AU) Wavelength (nm) 0.1 0.6 n =1.333 n =1.343 0 0.4 n =1.353 PIS PWS n =1.363 2 –0.1 0.2 Linear Linear n =1.373 580 600 620 640 660 1.33 1.34 1.35 1.36 1.37 1.33 1.34 1.35 1.36 1.37 Wavelength (nm) Refractive index (AU) Refractive index (AU) hi j PWV PIV Linear 1 Linear 628 0.9 1 626 0.8 0.3 0.8 0.8 0.7 624 626 628 630 0 200 400 0 200 400 0.2 Wavelength (nm) NP radius (nm) NP radius (nm) 0.6 r =50 nm 0.1 0.4 1 r =100 nm PWS r =250 nm PIS 0.2 r =500 nm Linear Linear –0.1 –1 580 600 620 640 660 0 200 400 0 200 400 Wavelength (nm) NP radius (nm) NP radius (nm) Fig. 2 Principle of peak intensity shift (PIS) and peak wavelength shift (PWS) on a PC surface. SEM images of a fabricated PC biosensor with a side views of the cross-section (inset: zoomed-in side view) and b top views (inset: zoomed-in top view). c FDTD simulation model of the PC surface (side view of the cross-section). d Simplified model as a waveguide on the PC surface (side view of the cross-section). e Normalized spectra with different background refractive indices (n = 1.333, 1.343, 1.353, 1.363, 1.373) on a PC surface without dielectric nanoparticles (inset: zoomed-in peak of the reflection spectra). Corresponding f peak intensity shift (PIS) (inset: peak intensity value (PIV)) and g peak wavelength shift (PWS) (inset: peak wavelength value (PWV)). h Normalized spectra with different sizes (radius of 50, 100, 250, 500 nm) of dielectric nanoparticles on the PC surface (inset: zoomed-in peak of the reflection spectra). Corresponding i PIS (inset: PIV) and j PWS (inset: PWV). Scale bar: 200 nm Reflectance (AU) Reflectance (AU) Reflectance (AU) Reflectance (AU) Normalized PIS (AU) Normalized PIS (AU) Normalized PIV (AU) Normalized PIV (AU) PWS (nm) PWS (nm) PWV (nm) PWV (nm) V Zhuo et al. Light: Science & Applications (2018) 7:9 Page 5 of 15 at the resonant wavelength (Fig. 1c) is the result of the Fabrication and preparation of the PC surface formation of surface-confined electromagnetic standing A room temperature replica molding approach is used waves that extend into the surrounding medium in the to fabricate the PC on a glass substrate using a quartz 53–80 form of an evanescent electromagnetic field . The mold template with a negative volume image of the photonic band gap of the PC strictly limits the lateral desired grating structure (fabricated with e-beam litho- propagation of light. Thus the PC exhibits a strong optical graphy and reactive ion etching). First, the quartz mold confinement of incident light into an infinitesimal volume template is thoroughly cleaned with a piranha solution (a that selectively interacts with surface-adsorbed cell com- mixture of sulfuric acid (H SO ) and hydrogen peroxide 2 4 ponents while being insensitive to the components of the (H O ), H SO :H O = 3:1) for approximately 3 hours to 2 2 2 4 2 2 cell body that are not engaged with the surface. Simula- remove organic residues from the surface of the master tions (Fig. 2) performed using the finite-difference time- template. The glass substrate is cleaned in an ultrasonic domain (FDTD) method show the spatial distribution of bath three times with isopropyl alcohol (IPA), acetone and the resonant electromagnetic field, which extends ~200 deionized (DI) water for 1 min in each solvent and then nm into the aqueous medium at the top of the PC. Pre- dried with nitrogen gas and treated with oxygen plasma. vious research has demonstrated that a specific location Second, the liquid UVCP is deposited between the quartz on the PC surface has a resonant reflected wavelength that mold template and glass substrate, and a high intensity can be independently measured from neighboring regions UV lamp is used to cure the liquid polymer to a solid and that the local PWV is determined by the dielectric state. After peeling the grating replica away from the permittivity of the biomaterial that is adsorbed at that quartz mold template, the nano-patterned surface is specific location . The PC surface can therefore act as a attached to a glass cover slip with an adhesive. Then PC proxy for a biological surface with a built-in capacity to fabrication is completed by reactive sputter deposition detect changes in the cell membrane components of cells (PVD 75, Kurt J. Lesker, Jefferson Hills, PA, USA) of a that attach to the PC within the evanescent field, pro- high refractive index thin film (TiO ) atop the grating viding a compelling platform for adhesion phenotyping of structure. Scanning electron microscopic (SEM) images of single cells (see Supplementary Materials Section S-1 for a cross-sectional view and a top view of the structure details). PC biosensor surfaces are inexpensively fabri- are shown in Fig. 2a, b, respectively. Next, before cell cated uniformly over large surface areas by a room tem- attachment experiments, the PC is cleaned in an ultra- perature nanoreplica molding process, as described sonic bath with IPA and DI water for 1 min each, followed 54,55 previously in refs. by drying with nitrogen gas. The PC is then treated , and are incorporated onto glass 49,50,52,81 microscope slides, described in refs. . with oxygen plasma to facilitate attachment of a liquid containment gasket formed from polydimethylsiloxane. Finally, the PC surface is hydrated with a phosphate- Modeling the PC surface for sensor design and simulation buffered saline solution and coated with a layer of A numerical electromagnetics simulation package ECM molecules (e.g., fibronectin) to promote cellular (FDTD, Lumerical Solutions, Inc., Vancouver, BC, attachment. Canada) is used to calculate the distribution of a resonant evanescent field on the PC biosensor surface. In our Photonic resonator outcoupler microscopy previous studies, the PC surface was modeled as an ideal The PROM instrument is a modified BF microscope 50,52 case with a rectangular nanostructure for simplicity . that uses a line-scanning approach to measure the spatial To more accurately represent the fabricated structure distribution of optical spectra across a PC surface with a (Fig. 2a, b), the model used in this study incorporates a submicron spatial resolution in the axial direction for 49,51 sidewall slope in a trapezoidal shape. As shown in Fig. 2c, label-free imaging (Fig. 1c) . An optical fiber-coupled d, the PC consists of a one-dimensional ultraviolet-cur- light-emitting diode is used as the light source, and a line- able polymer (UVCP) grating surface structure (refractive profiled (polarized perpendicular to the grating structure) index n = 1.46, grating depth d = 120 nm, period Λ = light beam illuminates the PC biosensor from below 0 g 400 nm, duty cycle f = 41.6%, sidewall angle θ = 85°) through a microscope objective lens (e.g., 10×). Illumi- g g coated with a thin film of TiO (refractive index n = 2.4, nation from below eliminates the possibility of the scat- 2 1 slab thickness d = 61 nm, duty cycle f = 50%, sidewall tering, absorption, and meniscus reflection and refraction s s angle θ = 82°) to generate a resonant reflected narrow- of materials in the cell media or cell body from effecting a band mode at a wavelength near λ = ~626 nm. The resonant reflected signal to the PC surface. The reflected adhesion of FAs on the PC surface is also modeled in light, containing the resonant reflected spectrum, passes FDTD, where the FA is represented as a homogeneous through the objective lens in the opposite direction and and lossless sphere (n = ~1.46, radius range 50–500 through the entrance slit of an imaging spectrometer and FA nm) composed of many protein molecules (Fig. 2e–j). is finally collected by a charge-coupled device camera, Zhuo et al. Light: Science & Applications (2018) 7:9 Page 6 of 15 which records the resonant reflected spectrum from each locally quench the PC resonance; (2) concentrated local pixel across the illuminated line on the PC surface. A high regions of high dielectric permittivity that can outcouple spatial resolution in the axial direction is obtained due to resonantly confined light by scattering . Interestingly, by the shallow evanescent field of the PC (~200 nm). The analyzing PROM data during cell attachment, we resolution in the lateral direction is determined by the observed the characteristics of PIV images that differ lateral propagation distance of resonant-coupled photons, substantially from PWS images. Although optical resulting in the detection of distinct surface-attached absorption at the PC resonant wavelength will efficiently objects for widely dispersed features at the micron size reduce the PIV in a highly localized manner, the protein scale . The same field of view can be re-scanned and lipid components of cells and cell membranes do not repeatedly to generate a sequence of images for the display strong absorption in the visible wavelength range. same cells. The current shortest available scan interval for Human tissues and live cells exhibit strong absorption in our instrument is ~10 s for ~100 × 100 µm , which is the infrared wavelength range; however, these wave- limited by the exposure time and speed of the motorized lengths are not utilized in our detection approach and scan stage. While characterizing the resolution perfor- thus are unlikely to be the dominant factors that con- mance through the intentional introduction of dielectric tribute to PIV reduction. Additionally, though metallic and metallic nanoparticles of a variety of sizes (30–500 elements comprise a small fraction of a cell’s atomic nm), we observe that dielectric objects not only induce a constituency, metal atoms are present as ions rather than shift in the PWV but also a reduction in the resonant peak as clusters that are capable of optical absorption in the intensity . When a cell attaches to the PC surface, the visible part of the spectrum. Scattering occurs when light peak resonant wavelength red-shifts from a lower wave- is forced to deviate from its original trajectory due to length (before cell attachment, e.g., λ = ~626 nm) to a localized non-uniformities in its propagation medium, BG higher wavelength (after cell attachment, e.g., λ = ~628 which occur, for example, when light propagating through cell nm). At the same time, the resonant reflection efficiency, water is reflected or refracted by a particle with a greater as measured by the peak intensity, changes from a higher refractive index. A highly concentrated region (e.g., FA PIV (before cell attachment, e.g., ~90% normalized to a cluster) with a greater refractive index than its sur- peak reflectance of 100% for the PC immersed in water) to roundings (e.g., cell media) generates more localized and a lower value (after cell attachment, e.g., 80% as a efficient scattering than a diffuse region with a gradual normalized PIV). A negative PIV shift indicates that gradient in the refractive index. Scattering effects also proteins in some areas of the cell bind to transmembrane become stronger when the size of a region with a proteins (e.g., integrins) to form more substantial FA refractive index contrast increases. Because the cross- clusters. The PROM instrument and sensor structure sectional area of a FA cluster is typically 0.2–1.0 µm ,we measure the resonant reflection characteristics of the can expect to observe measurable differences in the PC via the spectrum obtained from each ~0.6 × 0.6 µm scattering efficiency of membrane-associated protein pixel area, representing a total field of view region clusters as they form, change size, and subsequently dis- (e.g., ~300 × 300 µm ) of the PC surface. sipate. Light scattering from internal cell components, such as organelles and mitochondria, has recently been Stem cell culture utilized to achieve imaging contrast in the context of Murine dental epithelial stem cells (mHAT9a) changes that occur in precancerous cells that express were maintained in Dulbecco’s Modified Eagle phenotypic changes due to the expression of mutant 82–84 Medium supplemented with 10% fetal bovine serum genes . In PROM, we detect the modulation of and 1% Penicillin–Streptomycin. Stem cells were membrane-associated scattering that occurs due to FA cultured at a temperature of 37 °C and supplied formation by utilizing the ability of localized high with an environment of 5% CO humidified air during refractive index protein clusters to produce image imaging. contrast by reducing the reflection efficiency of a PC biosensor. Results and Discussion After analyzing the mechanism of PIV reduction, we Electromagnetic computer modeling of resonant study the useful cellular information that can be uniquely outcoupling from a PC by FA extracted from the measurement of this physical quantity. It is important to understand the physical mechanism Our hypothesis is that the dominant cause for the mea- that is responsible for the PIS in the context of cell sured PIV reduction is light scattering rather than attachment. Theoretical and experimental analyses sug- absorption. Thus a scattering model can represent the gest that a reduction in the PIV can occur by two interaction of light with a FA cluster since scattering mechanisms: (1) materials that act as efficient absorbers of describes the effect of an electromagnetic plane wave the resonant wavelength (such as gold nanoparticles) propagating through a dielectric particle. To predict the Zhuo et al. Light: Science & Applications (2018) 7:9 Page 7 of 15 effects on the resonant reflection spectrum from a PC that A difference of PWS and PIS images for the same cell at can be induced by a FA on its surface, we compared the each time point is clearly observed in the spectral data computed reflection spectra with and without a small acquired by PROM. In Fig. 3b, the dashed lines represent region of dielectric contrast on a PC surface. In a FDTD the background spectra for a representative pixel acquired electromagnetic computer model of a FA on a PC surface before cell attachment (0 min), and the dotted lines with (Fig. 2e–j), we represent the FA as a locus of a material dot markers represent the spectra of the same pixel after cell attachment (~16 min). PWS and PIS images are with a designated radius (50–500 nm) and designated refractive index elevation (n = ~1.46) in contrast to the simultaneously extracted from the spectra data at the FA surrounding medium (estimated to be n = ~1.333). The PWV and PIV, respectively (at every ~0.6 × 0.6 µm pixel simulation results demonstrate that, when the FA cluster area on the PC surface). In Fig. 3a, the regions with the is not present, the PWS in the resonant reflection spec- greatest values in the PWS and PIS images show distinct trum increases (Fig. 2g) as the refractive index contrast of distribution patterns, which suggests that these regions the surrounding media increases, whereas the PIS remains may represent two different physical mechanisms. For the same (Fig. 2f). However, when the FA cluster is pre- instance, the PWS image of the cell marked by the white sent, as the radius of the FA increases (Fig. 2h), the arrow (e.g., ~16 min) has a high PWS on the top and maximum in the resonant reflection spectrum decreases bottom of the cell body, whereas the PIS image of the (Fig. 2i, Inset), indicating that more energy is outcoupled same cell has a “ring” of a high PIS along the cell from the PC resonant standing wave. Thus the PWSs to a boundary. The zoomed-in images of the PWS and PIS higher wavelength (Fig. 2j), as expected, and the PIS taken 16 min after introduction of a single cell are shown (compared with the original PIS without a FA cluster) in Fig. 4a and S-Fig. 1, respectively, and the overlaid PWS increases as the FA cluster size increases (Fig. 2i) due to and PIS images (high intensity values only, PWS—green, the scattering-induced outcoupling (for more details, PIS—red, overlap—yellow) over the BF image of the same see Supplementary Materials S-2). cells show the distinct distribution patterns between these two physical quantities. The cross-sectional curves (L1 Dynamic PIS images of live cells and L2) along the diameter of the cell are plotted in Dynamic PIS images can be acquired by PROM during Fig. 4b, and the corresponding statistical results are live cell adhesion over an extended time period to create shown in Fig. 4c(N = 5 cells). The high intensity of the movies of FA development, which are difficult to acquire PWS (in green at the bottom of Fig. 4a) represents a via fluorescence imaging due to photobleaching. The PIS higher mass density of cellular materials associated with here is defined such that greater reductions in the the cell membrane. The high intensity of the PIS (in red at reflection efficiency are displayed as higher intensity the bottom of Fig. 4a) along the cell boundary represents values for a simpler visual comparison with other imaging the scattering outcoupling effect from the locally gener- modalities. As shown in Fig. 3a, the resulting PIS image ated FA clusters. Typically, a larger cluster size of protein sequence reveals that the cell periphery has a greater aggregates corresponds to a greater PIV reduction com- degree of scattering than the cell center. This “ring” effect pared with the background. Therefore, these measure- demonstrates that the PIS intensity is not homogeneously ments of local PWS and PIS can quantify the surface- distributed throughout the cell membrane. Fig. 3b shows attached cellular mass density and dimension of the FA spectra from three sample points marked in Fig. 3a(A— clusters dynamically and simultaneously. red, B—black, C—green) at different times. Initially, all three points are located outside of the cell (~0 min), and Comparison of PIS and fluorescence images all of the points show high resonant PIVs as highlighted Fluorescence images were acquired for the same cells to by the dashed line in the spectra (Fig. 3b). During cell further investigate the FA areas detected in the PIS ima- adhesion, the attachment perimeter expands and sur- ges. In Fig. 5a, the top row shows BF, PWS, and PIS passes the three points as the cell extends its attachment images and the bottom row displays fluorescence images area. Once the cell firmly attaches and adheres to the PC of the same cell with fluorescence tags applied to three surface (~16 min), points A’,B’, and C’ represent locations different cellular components (nucleus, actin, and vincu- inside, near, and outside of the cell boundary, respectively. lin). There is no obvious pattern similarity between the The solid lines in Fig. 3b demonstrate that the spectra of PIS image and actin (indicating the presence of cytoske- point A’ shifts to a lower resonance peak intensity, point leton components) or nucleus images. However, the B’ shifts to the lowest resonance peak intensity, and point fluorescent image of vinculin (a type of FA molecule) in C’ remains at the original resonance peak intensity. the bottom row of Fig. 5a indicates that filopodia reside in Considering all of the pixels within the cell, we observe a the FA area along the stem cell boundary. As shown in ring of enhanced scattering that encompasses much of the Fig. 5a in the right column, the PIS image shows a nearly cell periphery. identical distribution pattern along the cell peripheral Zhuo et al. Light: Science & Applications (2018) 7:9 Page 8 of 15 0 min 2 min 4 min 6 min 8 min 10 min 12 min 0.5 BF A 0.8 (AU) PIS 0.5 2.5 (nm) PWS 1.5 14 min 16 min 18 min 20 min 22 min 24 min 26 min 0.5 BF C’ B’ A’ 0.8 (AU) PIS 0.5 C’ B’ 2.5 (nm) A’ PWS 1.5 C’ 0.95 A BG-outside BG-inner cell A’ BG-cell boundary Outside Inner cell 0.8 Cell boundary B’ 0.6 0.7 0.4 624 626 628 630 Wavelength (nm) 0.2 600 620 640 660 Wavelength (nm) Fig. 3 PIV images and spectra during live stem cell (mHAT) attachment. a Sequence of PROM-acquired images at several time points (0–26 min) during the cell adhesion process. Top row: brightfield (BF) images; middle row: peak intensity shift (PIS) images; bottom row: peak wavelength shift (PWS) images. b PROM-measured spectra at three locations on the PC surface before (A, B, C, in dashed lines) and after (A’,B’,C’, in dotted lines) cell adhesion. Point A’ (red) represents the inner cell area; point B’ (black) represents the cell boundary; point C’ (green) represents the outside of the cell boundary. Scale bar: 20 μm region to that obtained by fluorescence microscopy with a (highlighted in the light-yellow regions), which indicates labeled vinculin where the FA areas are concentrated that the high intensity in the PIS images are probably co- along the cell boundary. Fig. 5b shows two cross-sections localized with the FA areas. Therefore, the dimensional along different radial directions (L3 and L4, as shown in change of the FA cluster can also be detected with PIS Fig. 5a) sampled across the cell diameter for the PWS images using PROM. (blue curves), PIS (red curves), and fluorescence-tagged Statistical analyses are shown in Fig. 5c(N = 5 cells) for images (nucleus—green, actin—magenta, vinculin— the fluorescent, PIS, and PWS images along the cell black). Both the PIS (in red) and vinculin (in black) curves boundary (marked as “Edge”) and within the nucleus area exhibit a similar “ring” effect along the cellular boundary (marked as “Inner”). Fig. 5c (Left) displays fluorescence Reflectance (AU) Reflectance (AU) Zhuo et al. Light: Science & Applications (2018) 7:9 Page 9 of 15 PWS (nm) PIS (AU) BF (AU) 2.5 0.8 1 L1 L1 L2 L2 0.5 1 0.5 Overlaid image Overlaid image (PWS+PIS) (PWS+PIS+BF) Cross-section L1 Cross-section L2 1 4 1 PWS PWS PIS PIS 0.5 2 0.5 0 0 0 0 10 20 30 40 50 10 20 30 40 50 Length (μm) Length (μm) 0.8 2.5 0.75 0.7 1.5 0.65 0.6 Edge Inner Edge Inner Fig. 4 Zoomed-in PROM images and cross-section (before and after mHAT cell adhesion). a PROM-acquired images (including PWS, PIS, and BF images) and their overlap images at ~16 min. b Comparison of cross-sectional curves (L1 and L2) of the PWS and PIS images. c Bar graph of the statistical comparison between the edge and the center of the cells. Scale bar: 20 μm images (including actin, nucleus, and vinculin) that scaffold protein, the distribution of actin is relatively demonstrate three different patterns of distribution along uniform, and thus the difference of the fluorescence the cell edge and center. The nucleus image only shows intensity between the cell edge and center in the actin high intensity within the area of the nucleus because the image is small. Vinculin is a membrane-cytoskeletal pro- fluorescent molecular probes only tag nucleic acid mate- tein that is often localized in the FA area because it par- rial, such as chromosomes. Actin mainly functions as a ticipates in the linkage between the transmembrane cytoskeleton molecule that is rapidly remodeled by protein (e.g., integrin) and cytoskeletal protein (e.g., actin). dynamically forming microfilaments to support the cell Therefore, a fluorescence dye for vinculin is often used to structure or participating in many important cellular visualize the locations of FAs. Fig. 5c clearly shows that processes, including cell division or cell signaling. As a the distribution of vinculin is mainly along the cell PWS (nm) Normalized PIS (AU) Normalized PIS (AU) PWS (nm) PWS (nm) Normalized PIS (AU) Zhuo et al. Light: Science & Applications (2018) 7:9 Page 10 of 15 PIS BF PWS L3 L3 L4 L4 1 2.5 (nm) 0.2 1 (AU) Nucleus Actin Vinculin L3 L3 L3 L4 L4 L4 Cross-section L3 Cross-section L4 Nucleus Nucleus Actin Actin Vinculin Vinculin 1 1 0.5 0.5 0 0 50 100 150 50 100 150 Length (μm) Length (μm) PWS PWS PIS PIS 0.5 0.5 50 100 150 50 100 150 Length (μm) Length (μm) Edge Inner 2.5 0.6 0.4 1.5 0.5 0.2 0.5 0 0 0 Actin Nucleus Vinculin Edge Inner Edge Inner Fig. 5 (See legend on next page.) Normalized intensity (AU) Normalized intensity (AU) Normalized intensity (AU) Label-free images Fluorescence images Normalized intensity (AU) Normalized intensity (AU) PWS (nm) Normalized PIS (AU) Zhuo et al. Light: Science & Applications (2018) 7:9 Page 11 of 15 (see figure on previous page) Fig. 5 Comparison of label-free images and fluorescence images with cross-section. a Top row: brightfield (BF), PWS, and PIS images; bottom row: fluorescence images, including dyes that selectively stain the nucleus, actin, and vinculin. b Comparison of the image cross-sections through lines (L3 and L4) for the fluorescent (nucleus—green, actin—magenta, vinculin—black), PWS (blue), and PIS (red) images. Light-yellow regions represent the regions near the cell edges, where the vinculin (black curves) and PIS (red curves) both have high intensities. c Statistical comparison among the fluorescence images (actin, nucleus, and vinculin) and label-free images (normalized PWS and PIS value comparison between cell edges and centers). Scale bar: 20 μm periphery for a surface-attached cell, which is highly (Fig. 6a). The mechanisms of the temporal curves for cell consistent with the distribution pattern of the PIS (high in adhesion for the PWS and PIS are shown in Fig. 6d, e (N “Edge”, low in “Inner”) (Fig. 5c Right), which is not similar = 5 cells). Directly comparing the dynamics between the to that of the PWS (Fig. 5c Middle). There is a small PWS and PIS is difficult because they have different difference between the PIS image and vinculin fluores- dynamic ranges. However, different slopes indicating cence image around (outside of) the nucleus. This is likely different increasing ratios along the temporal dimension because the vinculin fluorescence image is a transmission are observed if the PWS and normalized PIS are plotted image in the axial direction across the entire cell body together, as shown in Fig. 6f. (with a thickness of several microns to several tens of The PROM images show cell borders and intra-cell microns) including the nucleus area. By contrast, the PIS features that are approximately in the order of the pixel image is only measured through the evanescent field, size limit in the axial direction, which is near the dif- which has a thickness that is several hundred times fraction limit of the resonant wavelength. It is important thinner (several tens of nanometers) in the axial direction to put these images in the context of PCEM images that starting from the bottom of the cell body (before reaching were gathered previously from high-contrast objects. In nucleus). the lateral directions, although the effect of a point To highlight how information from PROM images dielectric object on the reflected wavelength from a PC complements those obtained by orthogonal imaging can extend to the surrounding pixels (because the electric modalities, five selected stem cells are shown in Fig. 6a field standing wave “samples” a greater lateral dimension imaged by PWS images, PIS images, confocal images with than one period), the outcoupling from a surface adsorbed fluorescence dyes (FL), phase-contrast images (PH), and scatterer (or absorber) is observed to be more highly SEM images. PROM images obtained using PWS and PIS localized. The full-widths at half-maximums of the point information reveal different features of cell attachment, spread function of TiO nanoparticles (e.g., diameter of and both show clear details of the cell attachment ~100 nm) were measured as ~1.20–1.56 µm for PWS boundary. PWS and PIS images highlight only behavior images and as ~0.95 µm for PIS images . Our previous associated with the cell–ECM interface and thus do not study shows that the spatial resolution of the dielectric show material in the upper cell body. Unlike SEM and objects in PWV-based PCEM images is directly correlated fluorescence images, PWS and PIS images yield dynamic with the refractive index contrast of the object. Although and highly quantitative information that can be visualized high-contrast objects (such as a TiO dielectric nano- graphically. As shown in Fig. 6b, c, the stem cell boundary particle or the edge of a photoresist pattern) can extend can be tracked along the local normal direction frame-by- their “influence” on the measured PWV by as much as a frame, enabling the PIS to be sampled along the cell couple of microns (e.g., 2–3 μm) in any direction, lower- boundary spatially and temporally at the same time. contrast objects have a much more limited perturbation. Associated dynamic analyses with different sampling In PROM images of attached cells, there is an extremely bands (S-Fig. 2, black for band 1—near cell boundary, low refractive index contrast between the attached cell white for band 2—inner region of the cell) in cells were membrane and surrounding cell media. Within the foot- performed on PWS images and PIS images. The resulting print of an attached cell, the refractive index contrast 2D maps shown in Fig. 6b, c represent the PWS and PIS between a FA and the neighboring cell membrane regions with spatiotemporal information along the cell boundary is even lower. These hypotheses were formed by the and time frames. Comparing the different bands between contrast observed at the cell attachment border and the both maps, the PIS increase is much higher in band 1 contrasting regions of attachment within a cell. Instead compared with that in band 2 (Fig. 6c), whereas the PWS of observing smeared borders that extend for several shows the opposite trend (Fig. 6b). The dramatic increase microns, we observed a contrast in PIS and PWS images of the PIS may be due to the aggregations of FAs along the with micron-scale features. These observations are con- cell boundaries, which is confirmed in the FL images sistent with our earlier measurements but have been Zhuo et al. Light: Science & Applications (2018) 7:9 Page 12 of 15 PWS PIS FL PH SEM SEM (zoom) Cell 1 Cell 2 5 μm Cell 3 Cell 4 Cell 5 0.6 0.4 0.8 (AU) 1.5 (nm) 20 μm bc PWS band 1 PWS band 2 PIS band 1 PIS band 2 (nm) (AU) 1.5 0.8 0.6 0.4 12 40 12 40 12 40 12 40 Time (min) Time (min) Time (min) Time (min) de 1.5 f PWS PIS 1.1 1 0.8 0.8 0.6 0.9 0.5 0.4 0.6 0.8 Band 1 0.2 Band 1 Band 2 Band 2 0 0 0.4 10 20 30 40 10 20 30 40 10 20 30 40 Time (min) Time (min) Time (min) Fig. 6 Image modality comparison and dynamic analysis of PROM images during cell adhesion. a Five selected cells imaged by peak wavelength shift (PWS), peak intensity shift (PIS), confocal fluorescence microscopy (FL) (red—actin, green—vinculin, blue—nucleus), phase-contrast microscopy (PH), and scanning electron microscopy (SEM). Scale bar: 20 μm. (Top-right inset: zoom-in SEM image for cell 1. Scale bar: 5 μm). 2D spatiotemporal maps are tracked from different bands (band 1—near cell boundary, band 2—inner region of cell) in stem cells with b PWS images and c normalized PIS images. Mean and standard deviation of the temporal curves of cell adhesion for d PWS, e normalized PIS, and f the comparison between PWS and normalized PIS Sampling window (AU) PWS (nm) Normalized PIS (AU) Sampling window (AU) PWS (nm) Normalized PIS (AU) Zhuo et al. Light: Science & Applications (2018) 7:9 Page 13 of 15 applied here for the first time in the context of attached University of Illinois Research Park, 1800 South Oak Street, Champaign, IL 61820, USA. Department of Electrical and Computer Engineering, University of cell images of the PIS. Illinois at Urbana-Champaign, Urbana, IL 61801, USA Conclusions Authors' contributions Y.Z. designed the experiments; Y.Z. and J.S.C. performed experiments and This study describes a label-free microscopic approach wrote the manuscript; Y.Z. and T.M. developed the analysis software and that quantitatively measures the scatter-induced changes performed data analysis; Y.Z. and H.Y. fabricated the PC sensors; B.T.C. and B.A. in the reflected intensity from a PC biosensor surface to H. provided guidance and edited the manuscript. reveal the kinetic evolution and spatial features of FAs Conflict of interest that form at the cell–surface interface. Compared to a The authors declare that they have no conflict of interest. sensing approach in which image contrast is generated by the dielectric permittivity of attached cell components, Supplementary information accompanies this paper at https://doi.org/ 10.1038/s41377-018-0001-5. PROM provides contrast in the reflected resonant inten- sity that is induced by the refractive index contrast of the Received: 11 May 2017 Revised: 13 February 2018 Accepted: 14 February localized protein clusters that occur at the cell–surface 2018 Accepted article preview online: 23 February 2018 interface, which comprise FA sites. Our hypothesis is supported by electromagnetic computer simulations that have modeled small and low refractive index contrast regions on a PC that induce measurable reductions in the References 1. Davies, P. F. & Tripathi, S. C. Mechanical stress mechanisms and the cell. An resonant reflection efficiency. Our hypothesis is also endothelial paradigm. Circ. Res. 72,239–245 (1993). supported by fluorescence microscopy of cells in which 2. Schaller, M. D. & Parsons, J. T. Focal adhesion kinase and associated proteins. the patterns of FA regions are similarly distributed as Curr. Opin. Cell Biol. 6,705–710 (1994). 3. Yamada, K.M.&Geiger,B.Molecular interactions in cell adhesion complexes. patterns of reflected intensity reduction measured by Curr. Opin. Cell Biol. 9,76–85 (1997). PROM. We show that images of the PIS and PWS can be 4. Pelham, R. J. Jr. & Wang, Y. Cell locomotion and focal adhesions are gathered from the same spectral information for the same regulated by substrate flexibility. Proc. Natl Acad. Sci. USA 94,13661–13665 (1997). cells and that the two imaging modalities have distinct 5. Turner, C. E. Paxillin and focal adhesion signalling. Nat. Cell Biol. 2,E231–E236 spatial patterns and thus provide complementary infor- (2000). mation about cell–surface activity. Dynamic images of the 6. Gerthoffer, W. T. & Gunst, S. J. Invited review: focal adhesion and small heat PIS and PWS can be repeatedly gathered over extended shock proteins in the regulation of actin remodeling and contractility in smooth muscle. J. Appl. Physiol. 91,963–972 (2001). time periods with a 10-s temporal resolution via a line- 7. Schaller, M. D. Biochemical signals and biological responses elicited by the scanning approach to generate time-course movies of focal adhesion kinase. Biochim. Biophys. Acta 1540,1–21 (2001). cell–surface behavior during processes that occur over 8. Wehrle-Haller, B. & Imhof,B.A.The inner lives of focal adhesions. Trends Cell Biol. 12, 382–389 (2002). several hours. As a label-free imaging approach, PROM 9. Chen,C.S., Alonso,J.L., Ostuni,E., Whitesides,G.M.&Ingber,D.E.Cellshape does not suffer from the limitations of fluorescence-based provides global control of focal adhesion assembly. Biochem. Biophys. Res microscopy, which include photobleaching and stain Commun. 307,355–361 (2003). 10. Carragher, N. O. & Frame, M. C. Focal adhesion and actin dynamics: a place cytotoxicity. We expect PROM to be a highly useful tool where kinases and proteases meet to promote invasion. Trends Cell Biol. 14, that can reveal the mechanisms of biological processes 241–249 (2004). that occur near the cell membrane when the membrane is 11. Owen, G. R.,Meredith, D. O.,apGwynn,I.&Richards, R.G.Focal adhesion quantification - a new assay of material biocompatibility? Review. Eur. Cell attached to ECM materials during cell migration, division, Mater. 9,85–96 (2005). discussion 85-96. metastasis, apoptosis, and stem cell differentiation. 12. Green,J.A.&Yamada,K.M.Three-dimensional microenvironmentsmodulate fibroblast signaling responses. Adv. Drug Deliv. Rev. 59,1293–1298 (2007). Acknowledgements 13. Gallant, N. D., Michael, K. E. & García, A. J. Cell adhesion strengthening: con- This work is supported by the National Science Foundation (NSF) Grant CBET tributions of adhesive area, integrin binding, and focal adhesion assembly. 11-32301 and National Institutes of Health (NIH) R01 DK099528 and NIH R21 Mol. Biol. Cell 16,4329–4340 (2005). EB018481. The content is solely the responsibility of the authors and does not 14. Wolfenson, H., Henis, Y. I., Geiger, B. & Bershadsky, A. D. The heel and toe of the necessarily represent the official views of the NSF and NIH. The authors would cell's foot: a multifaceted approach for understanding the structure and like to thank the Nano Sensor Groups (NSG), the staff at the Beckman Institute dynamics of focal adhesions. Cell Motil. Cytoskelet. 66,1017–1029 (2009). for Advanced Science and Technology, the Micro and Nanotechnology 15. Atilgan,E.&Ovryn, B. Nucleation and growth of integrin adhesions. Biophys. J. Laboratory (MNTL), the Institute for Genomic Biology (IGB), and the Center for 96,3555–3572 (2009). Innovative Instrumentation Technology (CiiT) at the University of Illinois at 16. Frisch, S. M., Vuori, K., Ruoslahti, E. & Chan-Hui, P. Y. Control of adhesion- Urbana-Champaign for their support. dependent cell survival by focal adhesion kinase. J. Cell Biol. 134,793–799 (1996). Author details 17. Schlaepfer, D. D.,Hauck,C.R. & Sieg,D.J.Signalingthrough focaladhesion Department of Bioengineering, University of Illinois at Urbana-Champaign, kinase. Prog.Biophys.Mol.Biol. 71, 435–478 (1999). Urbana, IL 61801, USA. Micro and Nanotechnology Laboratory, University of 18. Parsons, J.T., Martin,K. H., Slack, J. K.,Taylor, J. M. &Weed, S. A. Focaladhesion Illinois at Urbana-Champaign, Urbana, IL 61801, USA. Department of Chemical kinase: a regulator of focal adhesion dynamics and cell movement. Oncogene and Biomolecular Engineering, University of Illinois at Urbana-Champaign, 19,5606–5613 (2000). Urbana, IL 61801, USA. Carl R. Woese Institute for Genomic Biology, University 19. Petit, V. & Thiery, J. P. Focal adhesions: structure and dynamics. Biol. Cell 92, of Illinois at Urbana-Champaign, Urbana, IL 61801, USA. Atkins Building, 477–494 (2000). Zhuo et al. Light: Science & Applications (2018) 7:9 Page 14 of 15 20. Hauck,C.R., Hsia,D. A.& Schlaepfer,D. D.The focaladhesion 48. Berginski, M. E.,Vitriol, E.A., Hahn, K.M. & Gomez, S. M. High-resolution kinase--a regulator of cell migration and invasion. IUBMB Life 53,115–119 quantification of focal adhesion spatiotemporal dynamics in living cells. PLoS (2002). ONE 6, e22025 (2011). 21. Wozniak,M.A., Modzelewska, K.,Kwong,L.&Keely, P. J. Focaladhesion 49. Chen, W. et al. Photonic crystal enhanced microscopy for imaging of live cell regulation of cell behavior. Biochim. Biophys. Acta 1692,103–119 (2004). adhesion. Analyst 138,5886–5894 (2013). 22. McLean, G. W. et al. The role of focal-adhesion kinase in cancer - a new 50. Zhuo, Y. et al. Single nanoparticle detection using photonic crystal enhanced therapeutic opportunity. Nat. Rev. Cancer 5, 505–515 (2005). microscopy. Analyst 139,1007–1015 (2014). 23. Parsons, J. T., Horwitz, A. R. & Schwartz, M. A. Cell adhesion: integrating 51. Zhuo, Y. & Cunningham, B. T. Label-free biosensor imaging on photonic crystal cytoskeletal dynamics and cellular tension. Nat. Rev. Mol. Cell Biol. 11,633–643 surfaces. Sensors 15, 21613–21635 (2015). (2010). 52. Zhuo, Y. et al. Quantitative imaging of cell membrane-associated effective 24. Mitra, S. K., Hanson, D. A. & Schlaepfer, D. D. Focal adhesion kinase: in mass density using photonic crystal enhanced microscopy (PCEM). Prog. command and control of cell motility. Nat. Rev. Mol. Cell Biol. 6,56–68 Quantum Electron. 50,1–18 (2016). (2005). 53. Cunningham, B. T., Li, P., Lin, B. & Pepper, J. Colorimetric resonant reflection as 25. Damiano,J.S. & Dalton,W.S. Integrin-mediated drug resistance in multiple a direct biochemical assay technique. Sens. Actuators B Chem. 81,316–328 myeloma. Leuk. Lymphoma 38,71–81 (2000). (2002). 26. Hazlehurst, L. A.,Landowski,T.H.&Dalton,W.S.Roleofthe tumormicro- 54. Cunningham, B. T. et al. A plastic colorimetric resonant optical biosensor for environment in mediating de novo resistance to drugs and physiological multiparallel detection of label-free biochemical interactions. Sens. Actuators B mediators of cell death. Oncogene 22, 7396–7402 (2003). Chem. 85, 219–226 (2002). 27. McCulloch, C. A., Downey, G. P. & El-Gabalawy, H. Signalling platforms that 55. Cunningham, B. T. et al. Label-free assays on the BIND system. J. Biomol. Screen modulate the inflammatory response: new targets for drug development. Nat. 9,481–490 (2004). Rev. Drug Discov. 5,864–876 (2006). 56. Hessel, A. & Oliner, A. A. A new theory of Wood's anomalies on optical 28. Zhao, X. & Guan, J. L. Focal adhesion kinase and its signaling pathways in cell gratings. Appl. Opt. 4, 1275–1297 (1965). migration and angiogenesis. Adv. Drug Deliv. Rev. 63,610–615 (2011). 57. Yeh,P., Yariv, A. & Cho,A.Y.Optical surfacewaves in periodic layeredmedia. 29. Kim,D. H.&Wirtz, D. Focaladhesionsize uniquely predicts cell migration. Appl. Phys. Lett. 32,104–105 (1978). FASEB J. 27, 1351–1361 (2013). 58. Mashev, L. & Popov, E. Diffraction efficiency anomalies of multicoated 30. Kanchanawong, P. et al. Nanoscale architecture of integrin-based cell adhe- dielectric gratings. Opt. Commun. 51,131–136 (1984). sions. Nature 468,580–584 (2010). 59. Popov, E., Mashev, L. & Maystre, D. Theoretical study of the anomalies of 31. Geiger, B.,Spatz,J.P.&Bershadsky,A.D.Environmentalsensing through focal coated dielectric gratings. Opt. Acta 33,607–619 (1986). adhesions. Nat. Rev. Mol. Cell Biol. 10,21–33 (2009). 60. John, S. Strong localization of photons in certain disordered dielectric 32. Kusumi, A.,Tsunoyama,T.A., Hirosawa,K.M.&Kasai, R. S. &Fujiwara, T. K. superlattices. Phys.Rev.Lett. 58,2486–2489 (1987). Tracking single molecules at work in living cells. Nat. Chem. Biol. 10,524–532 61. Yablonovitch, E. Inhibited spontaneous emission in solid-state physics and (2014). electronics. Phys.Rev.Lett. 58,2059–2062 (1987). 33. Oakes, P.W.& Gardel, M.L.Stressing the limitsoffocal adhesion mechan- 62. Meade, R. D., Brommer, K. D.,Rappe,A. M.&Joannopoulos, J.D.Electro- osensitivity. Curr. Opin. Cell Biol. 30,68–73 (2014). magnetic Bloch waves at the surface of a photonic crystal. Phys.Rev.B 44, 34. Stehbens, S. J. & Wittmann, T. Analysis of focal adhesion turnover: a 10961–10964 (1991). quantitative live-cell imaging example. Methods Cell Biol. 123,335–346 63. Magnusson, R. & Wang, S. S. New principle for optical filters. Appl. Phys. Lett. 61, (2014). 1022–1024 (1992). 35. Deschout, H. et al. Complementarity of PALM and SOFI for super-resolution 64. Fan,S., Villeneuve,P.R., Joannopoulos,J.D.&Schubert,E.F.Highextraction live-cellimaging of focaladhesions. Nat. Commun. 7, 13693 (2016). efficiency of spontaneous emission from slabs of photonic crystals. Phys. Rev. 36. Maziveyi, M. & Alahari, S. K. Cell matrix adhesions in cancer: the proteins that Lett. 18,3294–3297 (1997). form the glue. Oncotarget 8, 48471–48487 (2017). 65. Joannopoulos, J. D., Villeneuve, P. R. & Fan, S. Photonic crystals: putting a new 37. Legant, W. R. et al. Multidimensional traction force microscopy reveals out-of- twist on light. Nature 386,143–149 (1997). plane rotational moments about focal adhesions. Proc. Natl. Acad. Sci. 110, 66. Kanskar, M. et al. Observation of leaky slab modes in an air-bridged semi- 881–886 (2013). conductor waveguide with a two-dimensional photonic lattice. Appl. Phys. 38. Colin-York, H. et al. Super-resolved traction force microscopy (STFM). Nano Lett. Lett. 70,1438–1440 (1997). 16,2633–2638 (2016). 67. Johnson, S. G., Fan, S., Villeneuve, P. R., Joannopoulos, J. D. & Kolodziejski, L. A. 39. Sarangi, B. R. et al. Coordination between intra- and extracellular forces reg- Guided modes in photonic crystal slabs. Phys. Rev. B 60, 5751–5758 (1999). ulates focal adhesion dynamics. Nano Lett. 17,399–406 (2017). 68. Boroditsky, M. et al. Spontaneous emission extraction and Purcell enhance- 40. Franz,C.M.& Muller,D.J. Analyzing focaladhesionstructure by atomic force ment from thin-film 2-D photonic crystals. J. Light Technol. 17,2096–2112 microscopy. J. Cell Sci. 118,5315–5323 (2005). (1999). 41. von Bilderling,C., Caldarola, M.,Masip,M.E., Bragas,A.V.&Pietrasanta, L. I. 69. Painter,O., Vuckovic,J.& Scherer, A. Defect modesofatwo-dimensional Monitoring in real-time focal adhesion protein dynamics in response to a photonic crystal in an optically thin dielectric slab. J. Opt. Soc. Am. B 16, discrete mechanical stimulus. Rev. Sci. Instrum. 88, 013703 (2017). 275–285 (1999). 42. Grashoff, C. et al. Measuring mechanical tension across vinculin reveals 70. Robertson, W. M. & May, M. S. Surface electromagnetic wave excitation on regulation of focal adhesion dynamics. Nature 466,263–266 (2010). one-dimensional photonic band gap arrays. Appl. Phys. Lett. 74,1800–1802 43. Figel, S. & Gelman, I. H. Focal adhesion kinase controls prostate cancer (1999). progression via intrinsic kinase and scaffolding functions. Anticancer Agents 71. Lin, S. Y., Chow, E., Johnson, S. G. & Joannopoulos, J. D. Demonstration of Med. Chem. 11, 607–616 (2011). highly efficient waveguiding in a photonic crystal slab at the 1.5-um wave- 44. Brooks, J., Watson, A. & Korcsmaros, T. Omics approaches to identify potential length. Opt. Lett. 25, 1297–1299 (2000). biomarkers of inflammatory diseases in the focal adhesion complex. Genomics 72. Pacradouni, V. et al. Photonic band structure of dielectric membranes peri- Proteomics Bioinformatics 15,101–109 (2017). odically textured in two dimensions. Phys. Rev. B 62, 4204–4207 (2000). 45. Reticker-Flynn,N.E.etal. Acombinatorial extracellular matrix platform iden- 73. Kuchinsky, S., Allan, D. C., Borrelli, N. F. & Cotteverte, J.-C. 3D localization in a tifies cell-extracellular matrix interactions that correlate with metastasis. Nat. channel waveguide in a photonic crystal with 2D periodicity. Opt. Commun. Commun. 3, 1122 (2012). 175,147–152 (2000). 46. Zhou, T., Marx, K. A., Dewilde, A. H., McIntosh, D. & Braunhut, S. J. Dynamic cell 74. Benisty, H. et al. Radiation losses of waveguide-based two-dimensional pho- adhesion and viscoelastic signatures distinguish normal from malignant tonic crystals: positive role of the substrate. Appl. Phys. Lett. 76,532–534 (2000). human mammary cells using quartz crystal microbalance. Anal. Biochem. 421, 75. Chutinan, A. & Noda, S. Waveguides and waveguide bends in two- dimensional photonic crystal slabs. Phys. Rev. B 62,4488–4492 (2000). 164–171 (2012). 47. Smolyakov, G. et al. Elasticity, adhesion, and tether extrusion on breast cancer 76. Joshi, B. et al. Phosphorylated caveolin-1 regulates Rho/ROCK-dependent focal cells provide a signature of their invasive potential. ACS Appl. Mater. Interfaces adhesion dynamics and tumor cell migration and invasion. Cancer Res. 68, 8, 27426–27431 (2016). 8210–8220 (2008). Zhuo et al. Light: Science & Applications (2018) 7:9 Page 15 of 15 77. Liu, J. N., Schulmerich, M. V., Bhargava, R. & Cunningham, B. T. Sculpting 81. Chen, W. L. et al. Enhanced live cell imaging via photonic narrowband Fano resonances inherent in the large-area mid-infrared photo- crystal enhanced fluorescence microscopy. Analyst 139,5954–5963 nic crystal microresonators for spectroscopic imaging. Opt. Express 22, (2014). 18142–18158 (2014). 82. Backman, V. et al. Polarized light scattering spectroscopy for quantitative 78. Chuang, S. L. Physics of Photonic Devices, 2nd edn (John Wiley & Sons Inc, New measurement of epithelial cellular structures in situ. IEEE J. Sel. Top. Quantum Jersey, USA, 2009). Electron. 5,1019–1026 (1999). 79. Foreman, M. Cavity Coupled Photonic Crystal Enhanced Fluorescence for High 83. Chandler, J. E., Cherkezyan, L., Subramanian, H. & Backman, V. Nanoscale Sensitivity Biomarker Detection. MSc thesis. University of Illinois at Urbana- refractive index fluctuations detected via sparse spectral microscopy. Biomed. Champaign (2016). Opt. Express 7,883–893 (2016). 80. Joannopoulos, J. D., Johnson, S. G., Winn, J. N. & Meade, R. D. Photonic Crystals: 84. Miao, Q.,Derbas, J.,Eid,A., Subramanian, H. &Backman, V.Automated Molding the Flow of Light, 2nd edn, (Princeton University Press, Princeton, cell selection using support vector machine for application to spectral 2008). nanocytology. Biomed. Res Int. 2016, 6090912 (2016). http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Light: Science & Applications Springer Journals

Quantitative analysis of focal adhesion dynamics using photonic resonator outcoupler microscopy (PROM)

Free
15 pages
Loading next page...
 
/lp/springer_journal/quantitative-analysis-of-focal-adhesion-dynamics-using-photonic-I8igeOdohA
Publisher
Nature Publishing Group UK
Copyright
Copyright © 2018 by The Author(s)
Subject
Physics; Physics, general; Applied and Technical Physics; Atomic, Molecular, Optical and Plasma Physics; Classical and Continuum Physics; Optics, Lasers, Photonics, Optical Devices
eISSN
2047-7538
D.O.I.
10.1038/s41377-018-0001-5
Publisher site
See Article on Publisher Site

Abstract

Focal adhesions are critical cell membrane components that regulate adhesion and migration and have cluster dimensions that correlate closely with adhesion engagement and migration speed. We utilized a label-free approach for dynamic, long-term, quantitative imaging of cell–surface interactions called photonic resonator outcoupler microscopy (PROM) in which membrane-associated protein aggregates outcoupled photons from the resonant evanescent field of a photonic crystal biosensor, resulting in a highly localized reduction of the reflected light intensity. By mapping the changes in the resonant reflected peak intensity from the biosensor surface, we demonstrate the ability of PROM to detect focal adhesion dimensions. Similar spatial distributions can be observed between PROM images and fluorescence-labeled images of focal adhesion areas in dental epithelial stem cells. In particular, we demonstrate that cell–surface contacts and focal adhesion formation can be imaged by two orthogonal label-free modalities in PROM simultaneously, providing a general-purpose tool for kinetic, high axial-resolution monitoring of cell interactions with basement membranes. and disassembly of a FA, the size of the FA cluster varies Introduction Focal adhesions (FAs), or cell–matrix adhesions, are and is highly correlated with the level of adhesion 13,29 large specialized proteins that are typically located at the engagement and migration speed . For example, non- interface between the cell membrane and extracellular mature focal complexes (FXs) are initially formed at the 1–24 matrix (ECM) (Fig. 1a, b) . FAs are critical for sup- leading edge of the cell (e.g., in the lamellipodia area) and porting the cell membrane structure and regulating signal are usually <0.2 µm . As the lamellipodia withdraws from transmission between the cytoskeleton (e.g., actin) and the leading edge, many FXs disassemble and release transmembrane receptors (e.g., integrins) during adhesion adhesion proteins back to the inner cell body, whereas 16–24 2 and migration . Monitoring the response of FA clus- some of the FXs grow larger (typically 1–10 µm ) and ters to drugs is one important mechanism by which the assemble into mature FA clusters by recruiting adapter 19,29 action of pharmaceutical compounds may be evaluated, proteins . Once the remaining FAs are in place, they particularly where approaches that enable characteriza- may form stationary attachment points by binding to the tion to be performed with a small number of cells are ECM, and a cell may utilize these anchors to migrate over 22,25–28 especially valuable . During the dynamic assembly the ECM by pushing and pulling the entire cellular 18,21,23 body . This insight into the dynamics of FA cluster formation and dissociation has been made possible by technical advances in the field of fluorescence and super Correspondence: Brian T. Cunningham (bcunning@illinois.edu) 1 30–36 Department of Bioengineering, University of Illinois at Urbana-Champaign, resolution microscopy . Optical modalities, including Urbana, IL 61801, USA total internal reflection fluorescence microscopy, Micro and Nanotechnology Laboratory, University of Illinois at Urbana- photoactivation localization microscopy (PALM), Champaign, Urbana, IL 61801, USA Full list of author information is available at the end of the article © The Author(s) 2018 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to theCreativeCommons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. 1234567890():,; 1234567890():,; 1234567890():,; 1234567890():,; Shift Zhuo et al. Light: Science & Applications (2018) 7:9 Page 2 of 15 ab Focal adhesion Evanescent field Extra cellular matrix Photonic crystal surface Photonic crystal surface Cell non-attached Cell attached Δλ (PWS) Cell BG ΔI (PIS) Photonic crystal Cell 0.9 biosensor 0.8 0.8 0.6 624 626 628 Objective lens 0.4 Wavelength (nm) Cylindrical 0.2 lens Dichroic 0 mirror 610 620 630 640 650 Wavelength (nm) Spectrometer Polarized LED beam splitter Mirror CCD camera Tube lens Fig. 1 Principle of the molecular-dynamics for cell attachment on a photonic crystal (PC) biosensor in photonic resonator outcoupler microscopy (PROM). Schematic representation of the molecular mechanism a before and b after a live cell attaches to the PC biosensor surface. c The principle of PROM imaging system. Inset: spectra shift before and after the cell attaches to the PC surface 17,37–39 stochastic optical reconstruction microscopy, and inter- micropillar substrates) , mechanical probing of cells 40,41 ferometric PALM, coupled with fluorescence tagging of (e.g., atomic force microscopy) , and single molecular the element(s) of FA clusters via administration of fluor- techniques (e.g., tension sensors) have allowed the escently labeled antibodies or incorporation of fluorescent quantification of molecular tension forces within FA reporter genes by transfection of cells, along with progress clusters as well as FA-mediated traction and adhesion made in single particle tracking algorithms, have allowed forces. researchers to quantify FA-associated parameters, such as Understanding the dynamics of FA formation and FA areas and sizes (x–y dimensions), FA architectures changes in FA-associated parameters is beneficial not only (x–y–z dimensions), FA turnover rates, and spatio- for understanding the fundamentals of biology but also temporal distributions of FA complexes. Additionally, for the field of biosensor diagnostics and screening for 36,43,44 developments in traction force measurements (e.g., based clinical applications . Changes in FA-associated on two-dimensional (2D) hydrogel substrates or parameters, such as FA sizes and traction forces, have Reflectance (AU) Reflectance (AU) Zhuo et al. Light: Science & Applications (2018) 7:9 Page 3 of 15 been linked to critical cellular processes, including value (PIV) from the PC before and after live cell metastasis, apoptosis, and chemotaxis, as well as pathol- attachment to acquire the peak intensity shift (PIS) at 9,29,36,45–47 ogies of cancers and other diseases . As such, each local voxel volume (Fig. 1a, b). Images of the PIS monitoring the response of FA clusters to drugs, for reveal highly localized and easily observable loci of protein example, is an important mechanism by which the action clusters that correlate with the spatial distribution and 22,25–28,36 of pharmaceutical compounds may be evaluated , size of FAs observed by fluorescence microscopy. We hypothesize that the observed reduction in reflected and high-throughput approaches that enable the char- acterization of small cell populations in real time are intensity from the PC is mainly caused by the outcoupling especially valuable for these applications. Currently of resonant standing wave photons via scattering. available techniques largely make use of fluorescence Because this imaging modality operates using an inde- tagging to mark individual FA proteins, which entails pendent sensing mechanism that obtains contrast through temporal limitations imposed by photobleaching and the formation of protein clusters at the cell–ECM inter- challenges associated with accurate quantitation and face that are capable of outcoupling light from the PC 9,11,48 long-term analysis . New tools are therefore required biosensor surface, we name this technique photonic to study the dynamic behavior of FA clusters and their resonator outcoupler microscopy (PROM) (Fig. 1c). In interaction with the ECM to characterize changes in FA our previous study, we report the first observation of a dynamics in live cells in situ. However, determining the reduced and highly localized reflected intensity in the dynamic activity of a FA cluster is challenging, especially context of nanoparticles with optical absorption at the with all of the FA proteins that are simultaneously active resonant wavelength of the PC . However, these obser- during the in situ assembly and disassembly processes in vations were made with very high contrast and highly live cells. Although a variety of approaches have been localized metal absorbers (for plasmonic nanoparticles) or utilized to investigate these processes, the detailed titanium dioxide (TiO ) nanoparticle dielectric scatterers. mechanism of FA assembly and disassembly in live cells, Although the reduced reflected resonant intensity from including the variability of the FA dimension, is poorly the high-contrast surface-attached TiO scatterers (the 9,11,48 understood . For instance, fluorescent tags are often refractive index of TiO nanoparticles (n = ~2.4) is 2 TiO2 used to mark individual FA proteins, but due to the much larger than that of the surrounding water medium temporal limitations imposed by photobleaching, accurate (n = ~1.333)) was the first observation with PROM, this quantitation and long-term analysis are exceedingly dif- study is the first to use PROM to observe scattered out- ficult to perform, whereas the cytotoxicity of fluorescent coupling from very low contrast FAs in live cell mem- tags compromises the viability of the cells under study. branes (averaged cell n = 1.35–1.38) to the surrounding cell Here we describe a label-free optical sensing approach medium (n = ~1.333). Using dental epithelial stem cells that combines two optical modalities to quantify the FA- attached to a fibronectin-coated ECM surface as a associated parameters that are critical for characterizing representative example, we demonstrate that PWS and spatiotemporal distribution patterns and the strength of PIS images of the same cells display distinct and com- FA clusters in real time. In our previous studies on cell plementary information. Although the regions with the imaging by photonic crystal enhanced microscopy greatest PWS are at the cell–surface interface in which (PCEM), we utilized an imaging modality in which the uniformly distributed regions with the greatest surface reflected resonant peak wavelength value (PWV) was engagement occur, the regions with the greatest PIS measured over the imaging field-of-view to derive images represent the formation of highly concentrated protein of the peak wavelength shift (PWS) that occur when cells clusters at the cell–surface interface that are capable of 49–52 attach to the photonic crystal (PC) surface . In these scattering photons. Therefore, we introduce PROM as a studies, we describe how the engagement of the cell quantitative, dynamic, and label-free approach to observe membrane components with the surface of the PC results the formation and evolution of FA cluster areas that are in highly localized shifts in the resonant reflected wave- otherwise challenging to observe with other available length from the biosensor , as well as the design of a imaging modalities, particularly for repeated observations modified brightfield (BF) microscope that enables visua- of the same cell population for extended time periods. lization of cell–surface attachments with a ~0.6 × 0.6 µm pixel size. The PWS image sequences clearly show the Materials and methods evolution of cell attachment through the engagement of PC biosensor the lipid bilayer cell membrane and internal cell- The PCs used in this study are subwavelength nanos- associated proteins within the ~200 nm deep evanescent tructured surfaces with a periodic modulation in the field region of the PC. In this study, we demonstrate a refractive index that acts as a narrow bandwidth resonant novel and orthogonal imaging modality within PCEM in optical reflector at one specific resonance wavelength 49,50,52–55 which we measure the resonant reflected peak intensity (Fig. 1) . The high reflection efficiency of the PC Zhuo et al. Light: Science & Applications (2018) 7:9 Page 4 of 15 ab c d s n d =d +d n g s (Medium) d g s n d =d 1 s d =d g  d =0 (Slab) d =0 (UVCP) 0 0 k z z TE TM n’ y n’ (Substrate) x e fg PIV PWV 1.02 Linear Linear 0.9 1 626 0.98 0.8 625 1.33 1.34 1.35 1.36 1.37 0.8 1.34 1.36 624 626 628 630 Refractive index (AU) Refractive index (AU) Wavelength (nm) 0.1 0.6 n =1.333 n =1.343 0 0.4 n =1.353 PIS PWS n =1.363 2 –0.1 0.2 Linear Linear n =1.373 580 600 620 640 660 1.33 1.34 1.35 1.36 1.37 1.33 1.34 1.35 1.36 1.37 Wavelength (nm) Refractive index (AU) Refractive index (AU) hi j PWV PIV Linear 1 Linear 628 0.9 1 626 0.8 0.3 0.8 0.8 0.7 624 626 628 630 0 200 400 0 200 400 0.2 Wavelength (nm) NP radius (nm) NP radius (nm) 0.6 r =50 nm 0.1 0.4 1 r =100 nm PWS r =250 nm PIS 0.2 r =500 nm Linear Linear –0.1 –1 580 600 620 640 660 0 200 400 0 200 400 Wavelength (nm) NP radius (nm) NP radius (nm) Fig. 2 Principle of peak intensity shift (PIS) and peak wavelength shift (PWS) on a PC surface. SEM images of a fabricated PC biosensor with a side views of the cross-section (inset: zoomed-in side view) and b top views (inset: zoomed-in top view). c FDTD simulation model of the PC surface (side view of the cross-section). d Simplified model as a waveguide on the PC surface (side view of the cross-section). e Normalized spectra with different background refractive indices (n = 1.333, 1.343, 1.353, 1.363, 1.373) on a PC surface without dielectric nanoparticles (inset: zoomed-in peak of the reflection spectra). Corresponding f peak intensity shift (PIS) (inset: peak intensity value (PIV)) and g peak wavelength shift (PWS) (inset: peak wavelength value (PWV)). h Normalized spectra with different sizes (radius of 50, 100, 250, 500 nm) of dielectric nanoparticles on the PC surface (inset: zoomed-in peak of the reflection spectra). Corresponding i PIS (inset: PIV) and j PWS (inset: PWV). Scale bar: 200 nm Reflectance (AU) Reflectance (AU) Reflectance (AU) Reflectance (AU) Normalized PIS (AU) Normalized PIS (AU) Normalized PIV (AU) Normalized PIV (AU) PWS (nm) PWS (nm) PWV (nm) PWV (nm) V Zhuo et al. Light: Science & Applications (2018) 7:9 Page 5 of 15 at the resonant wavelength (Fig. 1c) is the result of the Fabrication and preparation of the PC surface formation of surface-confined electromagnetic standing A room temperature replica molding approach is used waves that extend into the surrounding medium in the to fabricate the PC on a glass substrate using a quartz 53–80 form of an evanescent electromagnetic field . The mold template with a negative volume image of the photonic band gap of the PC strictly limits the lateral desired grating structure (fabricated with e-beam litho- propagation of light. Thus the PC exhibits a strong optical graphy and reactive ion etching). First, the quartz mold confinement of incident light into an infinitesimal volume template is thoroughly cleaned with a piranha solution (a that selectively interacts with surface-adsorbed cell com- mixture of sulfuric acid (H SO ) and hydrogen peroxide 2 4 ponents while being insensitive to the components of the (H O ), H SO :H O = 3:1) for approximately 3 hours to 2 2 2 4 2 2 cell body that are not engaged with the surface. Simula- remove organic residues from the surface of the master tions (Fig. 2) performed using the finite-difference time- template. The glass substrate is cleaned in an ultrasonic domain (FDTD) method show the spatial distribution of bath three times with isopropyl alcohol (IPA), acetone and the resonant electromagnetic field, which extends ~200 deionized (DI) water for 1 min in each solvent and then nm into the aqueous medium at the top of the PC. Pre- dried with nitrogen gas and treated with oxygen plasma. vious research has demonstrated that a specific location Second, the liquid UVCP is deposited between the quartz on the PC surface has a resonant reflected wavelength that mold template and glass substrate, and a high intensity can be independently measured from neighboring regions UV lamp is used to cure the liquid polymer to a solid and that the local PWV is determined by the dielectric state. After peeling the grating replica away from the permittivity of the biomaterial that is adsorbed at that quartz mold template, the nano-patterned surface is specific location . The PC surface can therefore act as a attached to a glass cover slip with an adhesive. Then PC proxy for a biological surface with a built-in capacity to fabrication is completed by reactive sputter deposition detect changes in the cell membrane components of cells (PVD 75, Kurt J. Lesker, Jefferson Hills, PA, USA) of a that attach to the PC within the evanescent field, pro- high refractive index thin film (TiO ) atop the grating viding a compelling platform for adhesion phenotyping of structure. Scanning electron microscopic (SEM) images of single cells (see Supplementary Materials Section S-1 for a cross-sectional view and a top view of the structure details). PC biosensor surfaces are inexpensively fabri- are shown in Fig. 2a, b, respectively. Next, before cell cated uniformly over large surface areas by a room tem- attachment experiments, the PC is cleaned in an ultra- perature nanoreplica molding process, as described sonic bath with IPA and DI water for 1 min each, followed 54,55 previously in refs. by drying with nitrogen gas. The PC is then treated , and are incorporated onto glass 49,50,52,81 microscope slides, described in refs. . with oxygen plasma to facilitate attachment of a liquid containment gasket formed from polydimethylsiloxane. Finally, the PC surface is hydrated with a phosphate- Modeling the PC surface for sensor design and simulation buffered saline solution and coated with a layer of A numerical electromagnetics simulation package ECM molecules (e.g., fibronectin) to promote cellular (FDTD, Lumerical Solutions, Inc., Vancouver, BC, attachment. Canada) is used to calculate the distribution of a resonant evanescent field on the PC biosensor surface. In our Photonic resonator outcoupler microscopy previous studies, the PC surface was modeled as an ideal The PROM instrument is a modified BF microscope 50,52 case with a rectangular nanostructure for simplicity . that uses a line-scanning approach to measure the spatial To more accurately represent the fabricated structure distribution of optical spectra across a PC surface with a (Fig. 2a, b), the model used in this study incorporates a submicron spatial resolution in the axial direction for 49,51 sidewall slope in a trapezoidal shape. As shown in Fig. 2c, label-free imaging (Fig. 1c) . An optical fiber-coupled d, the PC consists of a one-dimensional ultraviolet-cur- light-emitting diode is used as the light source, and a line- able polymer (UVCP) grating surface structure (refractive profiled (polarized perpendicular to the grating structure) index n = 1.46, grating depth d = 120 nm, period Λ = light beam illuminates the PC biosensor from below 0 g 400 nm, duty cycle f = 41.6%, sidewall angle θ = 85°) through a microscope objective lens (e.g., 10×). Illumi- g g coated with a thin film of TiO (refractive index n = 2.4, nation from below eliminates the possibility of the scat- 2 1 slab thickness d = 61 nm, duty cycle f = 50%, sidewall tering, absorption, and meniscus reflection and refraction s s angle θ = 82°) to generate a resonant reflected narrow- of materials in the cell media or cell body from effecting a band mode at a wavelength near λ = ~626 nm. The resonant reflected signal to the PC surface. The reflected adhesion of FAs on the PC surface is also modeled in light, containing the resonant reflected spectrum, passes FDTD, where the FA is represented as a homogeneous through the objective lens in the opposite direction and and lossless sphere (n = ~1.46, radius range 50–500 through the entrance slit of an imaging spectrometer and FA nm) composed of many protein molecules (Fig. 2e–j). is finally collected by a charge-coupled device camera, Zhuo et al. Light: Science & Applications (2018) 7:9 Page 6 of 15 which records the resonant reflected spectrum from each locally quench the PC resonance; (2) concentrated local pixel across the illuminated line on the PC surface. A high regions of high dielectric permittivity that can outcouple spatial resolution in the axial direction is obtained due to resonantly confined light by scattering . Interestingly, by the shallow evanescent field of the PC (~200 nm). The analyzing PROM data during cell attachment, we resolution in the lateral direction is determined by the observed the characteristics of PIV images that differ lateral propagation distance of resonant-coupled photons, substantially from PWS images. Although optical resulting in the detection of distinct surface-attached absorption at the PC resonant wavelength will efficiently objects for widely dispersed features at the micron size reduce the PIV in a highly localized manner, the protein scale . The same field of view can be re-scanned and lipid components of cells and cell membranes do not repeatedly to generate a sequence of images for the display strong absorption in the visible wavelength range. same cells. The current shortest available scan interval for Human tissues and live cells exhibit strong absorption in our instrument is ~10 s for ~100 × 100 µm , which is the infrared wavelength range; however, these wave- limited by the exposure time and speed of the motorized lengths are not utilized in our detection approach and scan stage. While characterizing the resolution perfor- thus are unlikely to be the dominant factors that con- mance through the intentional introduction of dielectric tribute to PIV reduction. Additionally, though metallic and metallic nanoparticles of a variety of sizes (30–500 elements comprise a small fraction of a cell’s atomic nm), we observe that dielectric objects not only induce a constituency, metal atoms are present as ions rather than shift in the PWV but also a reduction in the resonant peak as clusters that are capable of optical absorption in the intensity . When a cell attaches to the PC surface, the visible part of the spectrum. Scattering occurs when light peak resonant wavelength red-shifts from a lower wave- is forced to deviate from its original trajectory due to length (before cell attachment, e.g., λ = ~626 nm) to a localized non-uniformities in its propagation medium, BG higher wavelength (after cell attachment, e.g., λ = ~628 which occur, for example, when light propagating through cell nm). At the same time, the resonant reflection efficiency, water is reflected or refracted by a particle with a greater as measured by the peak intensity, changes from a higher refractive index. A highly concentrated region (e.g., FA PIV (before cell attachment, e.g., ~90% normalized to a cluster) with a greater refractive index than its sur- peak reflectance of 100% for the PC immersed in water) to roundings (e.g., cell media) generates more localized and a lower value (after cell attachment, e.g., 80% as a efficient scattering than a diffuse region with a gradual normalized PIV). A negative PIV shift indicates that gradient in the refractive index. Scattering effects also proteins in some areas of the cell bind to transmembrane become stronger when the size of a region with a proteins (e.g., integrins) to form more substantial FA refractive index contrast increases. Because the cross- clusters. The PROM instrument and sensor structure sectional area of a FA cluster is typically 0.2–1.0 µm ,we measure the resonant reflection characteristics of the can expect to observe measurable differences in the PC via the spectrum obtained from each ~0.6 × 0.6 µm scattering efficiency of membrane-associated protein pixel area, representing a total field of view region clusters as they form, change size, and subsequently dis- (e.g., ~300 × 300 µm ) of the PC surface. sipate. Light scattering from internal cell components, such as organelles and mitochondria, has recently been Stem cell culture utilized to achieve imaging contrast in the context of Murine dental epithelial stem cells (mHAT9a) changes that occur in precancerous cells that express were maintained in Dulbecco’s Modified Eagle phenotypic changes due to the expression of mutant 82–84 Medium supplemented with 10% fetal bovine serum genes . In PROM, we detect the modulation of and 1% Penicillin–Streptomycin. Stem cells were membrane-associated scattering that occurs due to FA cultured at a temperature of 37 °C and supplied formation by utilizing the ability of localized high with an environment of 5% CO humidified air during refractive index protein clusters to produce image imaging. contrast by reducing the reflection efficiency of a PC biosensor. Results and Discussion After analyzing the mechanism of PIV reduction, we Electromagnetic computer modeling of resonant study the useful cellular information that can be uniquely outcoupling from a PC by FA extracted from the measurement of this physical quantity. It is important to understand the physical mechanism Our hypothesis is that the dominant cause for the mea- that is responsible for the PIS in the context of cell sured PIV reduction is light scattering rather than attachment. Theoretical and experimental analyses sug- absorption. Thus a scattering model can represent the gest that a reduction in the PIV can occur by two interaction of light with a FA cluster since scattering mechanisms: (1) materials that act as efficient absorbers of describes the effect of an electromagnetic plane wave the resonant wavelength (such as gold nanoparticles) propagating through a dielectric particle. To predict the Zhuo et al. Light: Science & Applications (2018) 7:9 Page 7 of 15 effects on the resonant reflection spectrum from a PC that A difference of PWS and PIS images for the same cell at can be induced by a FA on its surface, we compared the each time point is clearly observed in the spectral data computed reflection spectra with and without a small acquired by PROM. In Fig. 3b, the dashed lines represent region of dielectric contrast on a PC surface. In a FDTD the background spectra for a representative pixel acquired electromagnetic computer model of a FA on a PC surface before cell attachment (0 min), and the dotted lines with (Fig. 2e–j), we represent the FA as a locus of a material dot markers represent the spectra of the same pixel after cell attachment (~16 min). PWS and PIS images are with a designated radius (50–500 nm) and designated refractive index elevation (n = ~1.46) in contrast to the simultaneously extracted from the spectra data at the FA surrounding medium (estimated to be n = ~1.333). The PWV and PIV, respectively (at every ~0.6 × 0.6 µm pixel simulation results demonstrate that, when the FA cluster area on the PC surface). In Fig. 3a, the regions with the is not present, the PWS in the resonant reflection spec- greatest values in the PWS and PIS images show distinct trum increases (Fig. 2g) as the refractive index contrast of distribution patterns, which suggests that these regions the surrounding media increases, whereas the PIS remains may represent two different physical mechanisms. For the same (Fig. 2f). However, when the FA cluster is pre- instance, the PWS image of the cell marked by the white sent, as the radius of the FA increases (Fig. 2h), the arrow (e.g., ~16 min) has a high PWS on the top and maximum in the resonant reflection spectrum decreases bottom of the cell body, whereas the PIS image of the (Fig. 2i, Inset), indicating that more energy is outcoupled same cell has a “ring” of a high PIS along the cell from the PC resonant standing wave. Thus the PWSs to a boundary. The zoomed-in images of the PWS and PIS higher wavelength (Fig. 2j), as expected, and the PIS taken 16 min after introduction of a single cell are shown (compared with the original PIS without a FA cluster) in Fig. 4a and S-Fig. 1, respectively, and the overlaid PWS increases as the FA cluster size increases (Fig. 2i) due to and PIS images (high intensity values only, PWS—green, the scattering-induced outcoupling (for more details, PIS—red, overlap—yellow) over the BF image of the same see Supplementary Materials S-2). cells show the distinct distribution patterns between these two physical quantities. The cross-sectional curves (L1 Dynamic PIS images of live cells and L2) along the diameter of the cell are plotted in Dynamic PIS images can be acquired by PROM during Fig. 4b, and the corresponding statistical results are live cell adhesion over an extended time period to create shown in Fig. 4c(N = 5 cells). The high intensity of the movies of FA development, which are difficult to acquire PWS (in green at the bottom of Fig. 4a) represents a via fluorescence imaging due to photobleaching. The PIS higher mass density of cellular materials associated with here is defined such that greater reductions in the the cell membrane. The high intensity of the PIS (in red at reflection efficiency are displayed as higher intensity the bottom of Fig. 4a) along the cell boundary represents values for a simpler visual comparison with other imaging the scattering outcoupling effect from the locally gener- modalities. As shown in Fig. 3a, the resulting PIS image ated FA clusters. Typically, a larger cluster size of protein sequence reveals that the cell periphery has a greater aggregates corresponds to a greater PIV reduction com- degree of scattering than the cell center. This “ring” effect pared with the background. Therefore, these measure- demonstrates that the PIS intensity is not homogeneously ments of local PWS and PIS can quantify the surface- distributed throughout the cell membrane. Fig. 3b shows attached cellular mass density and dimension of the FA spectra from three sample points marked in Fig. 3a(A— clusters dynamically and simultaneously. red, B—black, C—green) at different times. Initially, all three points are located outside of the cell (~0 min), and Comparison of PIS and fluorescence images all of the points show high resonant PIVs as highlighted Fluorescence images were acquired for the same cells to by the dashed line in the spectra (Fig. 3b). During cell further investigate the FA areas detected in the PIS ima- adhesion, the attachment perimeter expands and sur- ges. In Fig. 5a, the top row shows BF, PWS, and PIS passes the three points as the cell extends its attachment images and the bottom row displays fluorescence images area. Once the cell firmly attaches and adheres to the PC of the same cell with fluorescence tags applied to three surface (~16 min), points A’,B’, and C’ represent locations different cellular components (nucleus, actin, and vincu- inside, near, and outside of the cell boundary, respectively. lin). There is no obvious pattern similarity between the The solid lines in Fig. 3b demonstrate that the spectra of PIS image and actin (indicating the presence of cytoske- point A’ shifts to a lower resonance peak intensity, point leton components) or nucleus images. However, the B’ shifts to the lowest resonance peak intensity, and point fluorescent image of vinculin (a type of FA molecule) in C’ remains at the original resonance peak intensity. the bottom row of Fig. 5a indicates that filopodia reside in Considering all of the pixels within the cell, we observe a the FA area along the stem cell boundary. As shown in ring of enhanced scattering that encompasses much of the Fig. 5a in the right column, the PIS image shows a nearly cell periphery. identical distribution pattern along the cell peripheral Zhuo et al. Light: Science & Applications (2018) 7:9 Page 8 of 15 0 min 2 min 4 min 6 min 8 min 10 min 12 min 0.5 BF A 0.8 (AU) PIS 0.5 2.5 (nm) PWS 1.5 14 min 16 min 18 min 20 min 22 min 24 min 26 min 0.5 BF C’ B’ A’ 0.8 (AU) PIS 0.5 C’ B’ 2.5 (nm) A’ PWS 1.5 C’ 0.95 A BG-outside BG-inner cell A’ BG-cell boundary Outside Inner cell 0.8 Cell boundary B’ 0.6 0.7 0.4 624 626 628 630 Wavelength (nm) 0.2 600 620 640 660 Wavelength (nm) Fig. 3 PIV images and spectra during live stem cell (mHAT) attachment. a Sequence of PROM-acquired images at several time points (0–26 min) during the cell adhesion process. Top row: brightfield (BF) images; middle row: peak intensity shift (PIS) images; bottom row: peak wavelength shift (PWS) images. b PROM-measured spectra at three locations on the PC surface before (A, B, C, in dashed lines) and after (A’,B’,C’, in dotted lines) cell adhesion. Point A’ (red) represents the inner cell area; point B’ (black) represents the cell boundary; point C’ (green) represents the outside of the cell boundary. Scale bar: 20 μm region to that obtained by fluorescence microscopy with a (highlighted in the light-yellow regions), which indicates labeled vinculin where the FA areas are concentrated that the high intensity in the PIS images are probably co- along the cell boundary. Fig. 5b shows two cross-sections localized with the FA areas. Therefore, the dimensional along different radial directions (L3 and L4, as shown in change of the FA cluster can also be detected with PIS Fig. 5a) sampled across the cell diameter for the PWS images using PROM. (blue curves), PIS (red curves), and fluorescence-tagged Statistical analyses are shown in Fig. 5c(N = 5 cells) for images (nucleus—green, actin—magenta, vinculin— the fluorescent, PIS, and PWS images along the cell black). Both the PIS (in red) and vinculin (in black) curves boundary (marked as “Edge”) and within the nucleus area exhibit a similar “ring” effect along the cellular boundary (marked as “Inner”). Fig. 5c (Left) displays fluorescence Reflectance (AU) Reflectance (AU) Zhuo et al. Light: Science & Applications (2018) 7:9 Page 9 of 15 PWS (nm) PIS (AU) BF (AU) 2.5 0.8 1 L1 L1 L2 L2 0.5 1 0.5 Overlaid image Overlaid image (PWS+PIS) (PWS+PIS+BF) Cross-section L1 Cross-section L2 1 4 1 PWS PWS PIS PIS 0.5 2 0.5 0 0 0 0 10 20 30 40 50 10 20 30 40 50 Length (μm) Length (μm) 0.8 2.5 0.75 0.7 1.5 0.65 0.6 Edge Inner Edge Inner Fig. 4 Zoomed-in PROM images and cross-section (before and after mHAT cell adhesion). a PROM-acquired images (including PWS, PIS, and BF images) and their overlap images at ~16 min. b Comparison of cross-sectional curves (L1 and L2) of the PWS and PIS images. c Bar graph of the statistical comparison between the edge and the center of the cells. Scale bar: 20 μm images (including actin, nucleus, and vinculin) that scaffold protein, the distribution of actin is relatively demonstrate three different patterns of distribution along uniform, and thus the difference of the fluorescence the cell edge and center. The nucleus image only shows intensity between the cell edge and center in the actin high intensity within the area of the nucleus because the image is small. Vinculin is a membrane-cytoskeletal pro- fluorescent molecular probes only tag nucleic acid mate- tein that is often localized in the FA area because it par- rial, such as chromosomes. Actin mainly functions as a ticipates in the linkage between the transmembrane cytoskeleton molecule that is rapidly remodeled by protein (e.g., integrin) and cytoskeletal protein (e.g., actin). dynamically forming microfilaments to support the cell Therefore, a fluorescence dye for vinculin is often used to structure or participating in many important cellular visualize the locations of FAs. Fig. 5c clearly shows that processes, including cell division or cell signaling. As a the distribution of vinculin is mainly along the cell PWS (nm) Normalized PIS (AU) Normalized PIS (AU) PWS (nm) PWS (nm) Normalized PIS (AU) Zhuo et al. Light: Science & Applications (2018) 7:9 Page 10 of 15 PIS BF PWS L3 L3 L4 L4 1 2.5 (nm) 0.2 1 (AU) Nucleus Actin Vinculin L3 L3 L3 L4 L4 L4 Cross-section L3 Cross-section L4 Nucleus Nucleus Actin Actin Vinculin Vinculin 1 1 0.5 0.5 0 0 50 100 150 50 100 150 Length (μm) Length (μm) PWS PWS PIS PIS 0.5 0.5 50 100 150 50 100 150 Length (μm) Length (μm) Edge Inner 2.5 0.6 0.4 1.5 0.5 0.2 0.5 0 0 0 Actin Nucleus Vinculin Edge Inner Edge Inner Fig. 5 (See legend on next page.) Normalized intensity (AU) Normalized intensity (AU) Normalized intensity (AU) Label-free images Fluorescence images Normalized intensity (AU) Normalized intensity (AU) PWS (nm) Normalized PIS (AU) Zhuo et al. Light: Science & Applications (2018) 7:9 Page 11 of 15 (see figure on previous page) Fig. 5 Comparison of label-free images and fluorescence images with cross-section. a Top row: brightfield (BF), PWS, and PIS images; bottom row: fluorescence images, including dyes that selectively stain the nucleus, actin, and vinculin. b Comparison of the image cross-sections through lines (L3 and L4) for the fluorescent (nucleus—green, actin—magenta, vinculin—black), PWS (blue), and PIS (red) images. Light-yellow regions represent the regions near the cell edges, where the vinculin (black curves) and PIS (red curves) both have high intensities. c Statistical comparison among the fluorescence images (actin, nucleus, and vinculin) and label-free images (normalized PWS and PIS value comparison between cell edges and centers). Scale bar: 20 μm periphery for a surface-attached cell, which is highly (Fig. 6a). The mechanisms of the temporal curves for cell consistent with the distribution pattern of the PIS (high in adhesion for the PWS and PIS are shown in Fig. 6d, e (N “Edge”, low in “Inner”) (Fig. 5c Right), which is not similar = 5 cells). Directly comparing the dynamics between the to that of the PWS (Fig. 5c Middle). There is a small PWS and PIS is difficult because they have different difference between the PIS image and vinculin fluores- dynamic ranges. However, different slopes indicating cence image around (outside of) the nucleus. This is likely different increasing ratios along the temporal dimension because the vinculin fluorescence image is a transmission are observed if the PWS and normalized PIS are plotted image in the axial direction across the entire cell body together, as shown in Fig. 6f. (with a thickness of several microns to several tens of The PROM images show cell borders and intra-cell microns) including the nucleus area. By contrast, the PIS features that are approximately in the order of the pixel image is only measured through the evanescent field, size limit in the axial direction, which is near the dif- which has a thickness that is several hundred times fraction limit of the resonant wavelength. It is important thinner (several tens of nanometers) in the axial direction to put these images in the context of PCEM images that starting from the bottom of the cell body (before reaching were gathered previously from high-contrast objects. In nucleus). the lateral directions, although the effect of a point To highlight how information from PROM images dielectric object on the reflected wavelength from a PC complements those obtained by orthogonal imaging can extend to the surrounding pixels (because the electric modalities, five selected stem cells are shown in Fig. 6a field standing wave “samples” a greater lateral dimension imaged by PWS images, PIS images, confocal images with than one period), the outcoupling from a surface adsorbed fluorescence dyes (FL), phase-contrast images (PH), and scatterer (or absorber) is observed to be more highly SEM images. PROM images obtained using PWS and PIS localized. The full-widths at half-maximums of the point information reveal different features of cell attachment, spread function of TiO nanoparticles (e.g., diameter of and both show clear details of the cell attachment ~100 nm) were measured as ~1.20–1.56 µm for PWS boundary. PWS and PIS images highlight only behavior images and as ~0.95 µm for PIS images . Our previous associated with the cell–ECM interface and thus do not study shows that the spatial resolution of the dielectric show material in the upper cell body. Unlike SEM and objects in PWV-based PCEM images is directly correlated fluorescence images, PWS and PIS images yield dynamic with the refractive index contrast of the object. Although and highly quantitative information that can be visualized high-contrast objects (such as a TiO dielectric nano- graphically. As shown in Fig. 6b, c, the stem cell boundary particle or the edge of a photoresist pattern) can extend can be tracked along the local normal direction frame-by- their “influence” on the measured PWV by as much as a frame, enabling the PIS to be sampled along the cell couple of microns (e.g., 2–3 μm) in any direction, lower- boundary spatially and temporally at the same time. contrast objects have a much more limited perturbation. Associated dynamic analyses with different sampling In PROM images of attached cells, there is an extremely bands (S-Fig. 2, black for band 1—near cell boundary, low refractive index contrast between the attached cell white for band 2—inner region of the cell) in cells were membrane and surrounding cell media. Within the foot- performed on PWS images and PIS images. The resulting print of an attached cell, the refractive index contrast 2D maps shown in Fig. 6b, c represent the PWS and PIS between a FA and the neighboring cell membrane regions with spatiotemporal information along the cell boundary is even lower. These hypotheses were formed by the and time frames. Comparing the different bands between contrast observed at the cell attachment border and the both maps, the PIS increase is much higher in band 1 contrasting regions of attachment within a cell. Instead compared with that in band 2 (Fig. 6c), whereas the PWS of observing smeared borders that extend for several shows the opposite trend (Fig. 6b). The dramatic increase microns, we observed a contrast in PIS and PWS images of the PIS may be due to the aggregations of FAs along the with micron-scale features. These observations are con- cell boundaries, which is confirmed in the FL images sistent with our earlier measurements but have been Zhuo et al. Light: Science & Applications (2018) 7:9 Page 12 of 15 PWS PIS FL PH SEM SEM (zoom) Cell 1 Cell 2 5 μm Cell 3 Cell 4 Cell 5 0.6 0.4 0.8 (AU) 1.5 (nm) 20 μm bc PWS band 1 PWS band 2 PIS band 1 PIS band 2 (nm) (AU) 1.5 0.8 0.6 0.4 12 40 12 40 12 40 12 40 Time (min) Time (min) Time (min) Time (min) de 1.5 f PWS PIS 1.1 1 0.8 0.8 0.6 0.9 0.5 0.4 0.6 0.8 Band 1 0.2 Band 1 Band 2 Band 2 0 0 0.4 10 20 30 40 10 20 30 40 10 20 30 40 Time (min) Time (min) Time (min) Fig. 6 Image modality comparison and dynamic analysis of PROM images during cell adhesion. a Five selected cells imaged by peak wavelength shift (PWS), peak intensity shift (PIS), confocal fluorescence microscopy (FL) (red—actin, green—vinculin, blue—nucleus), phase-contrast microscopy (PH), and scanning electron microscopy (SEM). Scale bar: 20 μm. (Top-right inset: zoom-in SEM image for cell 1. Scale bar: 5 μm). 2D spatiotemporal maps are tracked from different bands (band 1—near cell boundary, band 2—inner region of cell) in stem cells with b PWS images and c normalized PIS images. Mean and standard deviation of the temporal curves of cell adhesion for d PWS, e normalized PIS, and f the comparison between PWS and normalized PIS Sampling window (AU) PWS (nm) Normalized PIS (AU) Sampling window (AU) PWS (nm) Normalized PIS (AU) Zhuo et al. Light: Science & Applications (2018) 7:9 Page 13 of 15 applied here for the first time in the context of attached University of Illinois Research Park, 1800 South Oak Street, Champaign, IL 61820, USA. Department of Electrical and Computer Engineering, University of cell images of the PIS. Illinois at Urbana-Champaign, Urbana, IL 61801, USA Conclusions Authors' contributions Y.Z. designed the experiments; Y.Z. and J.S.C. performed experiments and This study describes a label-free microscopic approach wrote the manuscript; Y.Z. and T.M. developed the analysis software and that quantitatively measures the scatter-induced changes performed data analysis; Y.Z. and H.Y. fabricated the PC sensors; B.T.C. and B.A. in the reflected intensity from a PC biosensor surface to H. provided guidance and edited the manuscript. reveal the kinetic evolution and spatial features of FAs Conflict of interest that form at the cell–surface interface. Compared to a The authors declare that they have no conflict of interest. sensing approach in which image contrast is generated by the dielectric permittivity of attached cell components, Supplementary information accompanies this paper at https://doi.org/ 10.1038/s41377-018-0001-5. PROM provides contrast in the reflected resonant inten- sity that is induced by the refractive index contrast of the Received: 11 May 2017 Revised: 13 February 2018 Accepted: 14 February localized protein clusters that occur at the cell–surface 2018 Accepted article preview online: 23 February 2018 interface, which comprise FA sites. Our hypothesis is supported by electromagnetic computer simulations that have modeled small and low refractive index contrast regions on a PC that induce measurable reductions in the References 1. Davies, P. F. & Tripathi, S. C. Mechanical stress mechanisms and the cell. An resonant reflection efficiency. Our hypothesis is also endothelial paradigm. Circ. Res. 72,239–245 (1993). supported by fluorescence microscopy of cells in which 2. Schaller, M. D. & Parsons, J. T. Focal adhesion kinase and associated proteins. the patterns of FA regions are similarly distributed as Curr. Opin. Cell Biol. 6,705–710 (1994). 3. Yamada, K.M.&Geiger,B.Molecular interactions in cell adhesion complexes. patterns of reflected intensity reduction measured by Curr. Opin. Cell Biol. 9,76–85 (1997). PROM. We show that images of the PIS and PWS can be 4. Pelham, R. J. Jr. & Wang, Y. Cell locomotion and focal adhesions are gathered from the same spectral information for the same regulated by substrate flexibility. Proc. Natl Acad. Sci. USA 94,13661–13665 (1997). cells and that the two imaging modalities have distinct 5. Turner, C. E. Paxillin and focal adhesion signalling. Nat. Cell Biol. 2,E231–E236 spatial patterns and thus provide complementary infor- (2000). mation about cell–surface activity. Dynamic images of the 6. Gerthoffer, W. T. & Gunst, S. J. Invited review: focal adhesion and small heat PIS and PWS can be repeatedly gathered over extended shock proteins in the regulation of actin remodeling and contractility in smooth muscle. J. Appl. Physiol. 91,963–972 (2001). time periods with a 10-s temporal resolution via a line- 7. Schaller, M. D. Biochemical signals and biological responses elicited by the scanning approach to generate time-course movies of focal adhesion kinase. Biochim. Biophys. Acta 1540,1–21 (2001). cell–surface behavior during processes that occur over 8. Wehrle-Haller, B. & Imhof,B.A.The inner lives of focal adhesions. Trends Cell Biol. 12, 382–389 (2002). several hours. As a label-free imaging approach, PROM 9. Chen,C.S., Alonso,J.L., Ostuni,E., Whitesides,G.M.&Ingber,D.E.Cellshape does not suffer from the limitations of fluorescence-based provides global control of focal adhesion assembly. Biochem. Biophys. Res microscopy, which include photobleaching and stain Commun. 307,355–361 (2003). 10. Carragher, N. O. & Frame, M. C. Focal adhesion and actin dynamics: a place cytotoxicity. We expect PROM to be a highly useful tool where kinases and proteases meet to promote invasion. Trends Cell Biol. 14, that can reveal the mechanisms of biological processes 241–249 (2004). that occur near the cell membrane when the membrane is 11. Owen, G. R.,Meredith, D. O.,apGwynn,I.&Richards, R.G.Focal adhesion quantification - a new assay of material biocompatibility? Review. Eur. Cell attached to ECM materials during cell migration, division, Mater. 9,85–96 (2005). discussion 85-96. metastasis, apoptosis, and stem cell differentiation. 12. Green,J.A.&Yamada,K.M.Three-dimensional microenvironmentsmodulate fibroblast signaling responses. Adv. Drug Deliv. Rev. 59,1293–1298 (2007). Acknowledgements 13. Gallant, N. D., Michael, K. E. & García, A. J. Cell adhesion strengthening: con- This work is supported by the National Science Foundation (NSF) Grant CBET tributions of adhesive area, integrin binding, and focal adhesion assembly. 11-32301 and National Institutes of Health (NIH) R01 DK099528 and NIH R21 Mol. Biol. Cell 16,4329–4340 (2005). EB018481. The content is solely the responsibility of the authors and does not 14. Wolfenson, H., Henis, Y. I., Geiger, B. & Bershadsky, A. D. The heel and toe of the necessarily represent the official views of the NSF and NIH. The authors would cell's foot: a multifaceted approach for understanding the structure and like to thank the Nano Sensor Groups (NSG), the staff at the Beckman Institute dynamics of focal adhesions. Cell Motil. Cytoskelet. 66,1017–1029 (2009). for Advanced Science and Technology, the Micro and Nanotechnology 15. Atilgan,E.&Ovryn, B. Nucleation and growth of integrin adhesions. Biophys. J. Laboratory (MNTL), the Institute for Genomic Biology (IGB), and the Center for 96,3555–3572 (2009). Innovative Instrumentation Technology (CiiT) at the University of Illinois at 16. Frisch, S. M., Vuori, K., Ruoslahti, E. & Chan-Hui, P. Y. Control of adhesion- Urbana-Champaign for their support. dependent cell survival by focal adhesion kinase. J. Cell Biol. 134,793–799 (1996). Author details 17. Schlaepfer, D. D.,Hauck,C.R. & Sieg,D.J.Signalingthrough focaladhesion Department of Bioengineering, University of Illinois at Urbana-Champaign, kinase. Prog.Biophys.Mol.Biol. 71, 435–478 (1999). Urbana, IL 61801, USA. Micro and Nanotechnology Laboratory, University of 18. Parsons, J.T., Martin,K. H., Slack, J. K.,Taylor, J. M. &Weed, S. A. Focaladhesion Illinois at Urbana-Champaign, Urbana, IL 61801, USA. Department of Chemical kinase: a regulator of focal adhesion dynamics and cell movement. Oncogene and Biomolecular Engineering, University of Illinois at Urbana-Champaign, 19,5606–5613 (2000). Urbana, IL 61801, USA. Carl R. Woese Institute for Genomic Biology, University 19. Petit, V. & Thiery, J. P. Focal adhesions: structure and dynamics. Biol. Cell 92, of Illinois at Urbana-Champaign, Urbana, IL 61801, USA. Atkins Building, 477–494 (2000). Zhuo et al. Light: Science & Applications (2018) 7:9 Page 14 of 15 20. Hauck,C.R., Hsia,D. A.& Schlaepfer,D. D.The focaladhesion 48. Berginski, M. E.,Vitriol, E.A., Hahn, K.M. & Gomez, S. M. High-resolution kinase--a regulator of cell migration and invasion. IUBMB Life 53,115–119 quantification of focal adhesion spatiotemporal dynamics in living cells. PLoS (2002). ONE 6, e22025 (2011). 21. Wozniak,M.A., Modzelewska, K.,Kwong,L.&Keely, P. J. Focaladhesion 49. Chen, W. et al. Photonic crystal enhanced microscopy for imaging of live cell regulation of cell behavior. Biochim. Biophys. Acta 1692,103–119 (2004). adhesion. Analyst 138,5886–5894 (2013). 22. McLean, G. W. et al. The role of focal-adhesion kinase in cancer - a new 50. Zhuo, Y. et al. Single nanoparticle detection using photonic crystal enhanced therapeutic opportunity. Nat. Rev. Cancer 5, 505–515 (2005). microscopy. Analyst 139,1007–1015 (2014). 23. Parsons, J. T., Horwitz, A. R. & Schwartz, M. A. Cell adhesion: integrating 51. Zhuo, Y. & Cunningham, B. T. Label-free biosensor imaging on photonic crystal cytoskeletal dynamics and cellular tension. Nat. Rev. Mol. Cell Biol. 11,633–643 surfaces. Sensors 15, 21613–21635 (2015). (2010). 52. Zhuo, Y. et al. Quantitative imaging of cell membrane-associated effective 24. Mitra, S. K., Hanson, D. A. & Schlaepfer, D. D. Focal adhesion kinase: in mass density using photonic crystal enhanced microscopy (PCEM). Prog. command and control of cell motility. Nat. Rev. Mol. Cell Biol. 6,56–68 Quantum Electron. 50,1–18 (2016). (2005). 53. Cunningham, B. T., Li, P., Lin, B. & Pepper, J. Colorimetric resonant reflection as 25. Damiano,J.S. & Dalton,W.S. Integrin-mediated drug resistance in multiple a direct biochemical assay technique. Sens. Actuators B Chem. 81,316–328 myeloma. Leuk. Lymphoma 38,71–81 (2000). (2002). 26. Hazlehurst, L. A.,Landowski,T.H.&Dalton,W.S.Roleofthe tumormicro- 54. Cunningham, B. T. et al. A plastic colorimetric resonant optical biosensor for environment in mediating de novo resistance to drugs and physiological multiparallel detection of label-free biochemical interactions. Sens. Actuators B mediators of cell death. Oncogene 22, 7396–7402 (2003). Chem. 85, 219–226 (2002). 27. McCulloch, C. A., Downey, G. P. & El-Gabalawy, H. Signalling platforms that 55. Cunningham, B. T. et al. Label-free assays on the BIND system. J. Biomol. Screen modulate the inflammatory response: new targets for drug development. Nat. 9,481–490 (2004). Rev. Drug Discov. 5,864–876 (2006). 56. Hessel, A. & Oliner, A. A. A new theory of Wood's anomalies on optical 28. Zhao, X. & Guan, J. L. Focal adhesion kinase and its signaling pathways in cell gratings. Appl. Opt. 4, 1275–1297 (1965). migration and angiogenesis. Adv. Drug Deliv. Rev. 63,610–615 (2011). 57. Yeh,P., Yariv, A. & Cho,A.Y.Optical surfacewaves in periodic layeredmedia. 29. Kim,D. H.&Wirtz, D. Focaladhesionsize uniquely predicts cell migration. Appl. Phys. Lett. 32,104–105 (1978). FASEB J. 27, 1351–1361 (2013). 58. Mashev, L. & Popov, E. Diffraction efficiency anomalies of multicoated 30. Kanchanawong, P. et al. Nanoscale architecture of integrin-based cell adhe- dielectric gratings. Opt. Commun. 51,131–136 (1984). sions. Nature 468,580–584 (2010). 59. Popov, E., Mashev, L. & Maystre, D. Theoretical study of the anomalies of 31. Geiger, B.,Spatz,J.P.&Bershadsky,A.D.Environmentalsensing through focal coated dielectric gratings. Opt. Acta 33,607–619 (1986). adhesions. Nat. Rev. Mol. Cell Biol. 10,21–33 (2009). 60. John, S. Strong localization of photons in certain disordered dielectric 32. Kusumi, A.,Tsunoyama,T.A., Hirosawa,K.M.&Kasai, R. S. &Fujiwara, T. K. superlattices. Phys.Rev.Lett. 58,2486–2489 (1987). Tracking single molecules at work in living cells. Nat. Chem. Biol. 10,524–532 61. Yablonovitch, E. Inhibited spontaneous emission in solid-state physics and (2014). electronics. Phys.Rev.Lett. 58,2059–2062 (1987). 33. Oakes, P.W.& Gardel, M.L.Stressing the limitsoffocal adhesion mechan- 62. Meade, R. D., Brommer, K. D.,Rappe,A. M.&Joannopoulos, J.D.Electro- osensitivity. Curr. Opin. Cell Biol. 30,68–73 (2014). magnetic Bloch waves at the surface of a photonic crystal. Phys.Rev.B 44, 34. Stehbens, S. J. & Wittmann, T. Analysis of focal adhesion turnover: a 10961–10964 (1991). quantitative live-cell imaging example. Methods Cell Biol. 123,335–346 63. Magnusson, R. & Wang, S. S. New principle for optical filters. Appl. Phys. Lett. 61, (2014). 1022–1024 (1992). 35. Deschout, H. et al. Complementarity of PALM and SOFI for super-resolution 64. Fan,S., Villeneuve,P.R., Joannopoulos,J.D.&Schubert,E.F.Highextraction live-cellimaging of focaladhesions. Nat. Commun. 7, 13693 (2016). efficiency of spontaneous emission from slabs of photonic crystals. Phys. Rev. 36. Maziveyi, M. & Alahari, S. K. Cell matrix adhesions in cancer: the proteins that Lett. 18,3294–3297 (1997). form the glue. Oncotarget 8, 48471–48487 (2017). 65. Joannopoulos, J. D., Villeneuve, P. R. & Fan, S. Photonic crystals: putting a new 37. Legant, W. R. et al. Multidimensional traction force microscopy reveals out-of- twist on light. Nature 386,143–149 (1997). plane rotational moments about focal adhesions. Proc. Natl. Acad. Sci. 110, 66. Kanskar, M. et al. Observation of leaky slab modes in an air-bridged semi- 881–886 (2013). conductor waveguide with a two-dimensional photonic lattice. Appl. Phys. 38. Colin-York, H. et al. Super-resolved traction force microscopy (STFM). Nano Lett. Lett. 70,1438–1440 (1997). 16,2633–2638 (2016). 67. Johnson, S. G., Fan, S., Villeneuve, P. R., Joannopoulos, J. D. & Kolodziejski, L. A. 39. Sarangi, B. R. et al. Coordination between intra- and extracellular forces reg- Guided modes in photonic crystal slabs. Phys. Rev. B 60, 5751–5758 (1999). ulates focal adhesion dynamics. Nano Lett. 17,399–406 (2017). 68. Boroditsky, M. et al. Spontaneous emission extraction and Purcell enhance- 40. Franz,C.M.& Muller,D.J. Analyzing focaladhesionstructure by atomic force ment from thin-film 2-D photonic crystals. J. Light Technol. 17,2096–2112 microscopy. J. Cell Sci. 118,5315–5323 (2005). (1999). 41. von Bilderling,C., Caldarola, M.,Masip,M.E., Bragas,A.V.&Pietrasanta, L. I. 69. Painter,O., Vuckovic,J.& Scherer, A. Defect modesofatwo-dimensional Monitoring in real-time focal adhesion protein dynamics in response to a photonic crystal in an optically thin dielectric slab. J. Opt. Soc. Am. B 16, discrete mechanical stimulus. Rev. Sci. Instrum. 88, 013703 (2017). 275–285 (1999). 42. Grashoff, C. et al. Measuring mechanical tension across vinculin reveals 70. Robertson, W. M. & May, M. S. Surface electromagnetic wave excitation on regulation of focal adhesion dynamics. Nature 466,263–266 (2010). one-dimensional photonic band gap arrays. Appl. Phys. Lett. 74,1800–1802 43. Figel, S. & Gelman, I. H. Focal adhesion kinase controls prostate cancer (1999). progression via intrinsic kinase and scaffolding functions. Anticancer Agents 71. Lin, S. Y., Chow, E., Johnson, S. G. & Joannopoulos, J. D. Demonstration of Med. Chem. 11, 607–616 (2011). highly efficient waveguiding in a photonic crystal slab at the 1.5-um wave- 44. Brooks, J., Watson, A. & Korcsmaros, T. Omics approaches to identify potential length. Opt. Lett. 25, 1297–1299 (2000). biomarkers of inflammatory diseases in the focal adhesion complex. Genomics 72. Pacradouni, V. et al. Photonic band structure of dielectric membranes peri- Proteomics Bioinformatics 15,101–109 (2017). odically textured in two dimensions. Phys. Rev. B 62, 4204–4207 (2000). 45. Reticker-Flynn,N.E.etal. Acombinatorial extracellular matrix platform iden- 73. Kuchinsky, S., Allan, D. C., Borrelli, N. F. & Cotteverte, J.-C. 3D localization in a tifies cell-extracellular matrix interactions that correlate with metastasis. Nat. channel waveguide in a photonic crystal with 2D periodicity. Opt. Commun. Commun. 3, 1122 (2012). 175,147–152 (2000). 46. Zhou, T., Marx, K. A., Dewilde, A. H., McIntosh, D. & Braunhut, S. J. Dynamic cell 74. Benisty, H. et al. Radiation losses of waveguide-based two-dimensional pho- adhesion and viscoelastic signatures distinguish normal from malignant tonic crystals: positive role of the substrate. Appl. Phys. Lett. 76,532–534 (2000). human mammary cells using quartz crystal microbalance. Anal. Biochem. 421, 75. Chutinan, A. & Noda, S. Waveguides and waveguide bends in two- dimensional photonic crystal slabs. Phys. Rev. B 62,4488–4492 (2000). 164–171 (2012). 47. Smolyakov, G. et al. Elasticity, adhesion, and tether extrusion on breast cancer 76. Joshi, B. et al. Phosphorylated caveolin-1 regulates Rho/ROCK-dependent focal cells provide a signature of their invasive potential. ACS Appl. Mater. Interfaces adhesion dynamics and tumor cell migration and invasion. Cancer Res. 68, 8, 27426–27431 (2016). 8210–8220 (2008). Zhuo et al. Light: Science & Applications (2018) 7:9 Page 15 of 15 77. Liu, J. N., Schulmerich, M. V., Bhargava, R. & Cunningham, B. T. Sculpting 81. Chen, W. L. et al. Enhanced live cell imaging via photonic narrowband Fano resonances inherent in the large-area mid-infrared photo- crystal enhanced fluorescence microscopy. Analyst 139,5954–5963 nic crystal microresonators for spectroscopic imaging. Opt. Express 22, (2014). 18142–18158 (2014). 82. Backman, V. et al. Polarized light scattering spectroscopy for quantitative 78. Chuang, S. L. Physics of Photonic Devices, 2nd edn (John Wiley & Sons Inc, New measurement of epithelial cellular structures in situ. IEEE J. Sel. Top. Quantum Jersey, USA, 2009). Electron. 5,1019–1026 (1999). 79. Foreman, M. Cavity Coupled Photonic Crystal Enhanced Fluorescence for High 83. Chandler, J. E., Cherkezyan, L., Subramanian, H. & Backman, V. Nanoscale Sensitivity Biomarker Detection. MSc thesis. University of Illinois at Urbana- refractive index fluctuations detected via sparse spectral microscopy. Biomed. Champaign (2016). Opt. Express 7,883–893 (2016). 80. Joannopoulos, J. D., Johnson, S. G., Winn, J. N. & Meade, R. D. Photonic Crystals: 84. Miao, Q.,Derbas, J.,Eid,A., Subramanian, H. &Backman, V.Automated Molding the Flow of Light, 2nd edn, (Princeton University Press, Princeton, cell selection using support vector machine for application to spectral 2008). nanocytology. Biomed. Res Int. 2016, 6090912 (2016).

Journal

Light: Science & ApplicationsSpringer Journals

Published: May 30, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off