The generation of cancer cells from a non-tumorous state and their conversion to the metastatic state are of critical importance in cancer cell biology and medicine. The classification of cancer cells from non-cancer cells and into distinct subtypes has proven instrumental in tailoring therapeutic strategies and predicting patient prognosis. Cancer cells have been characterized in terms of their biological (i.e., genomic profile [1], transcriptional profile [2], metastatic ability [3]) and physical properties (i.e., morphology), viscoelasticity [4], adhesion strength [5], migration [6], and receptor dynamics [7]). However, the connection between cell properties and cellular phenotypes, typically metastatic potential, remains unclear. Understanding cancer cell subtypes, especially those associated with metastatic propensity, holds promise for refining prognostic assessments and guiding treatment approaches.

Cancer metastasis, the process by which cancer cells disseminate from the primary tumor to establish secondary growth in distant organs, is a complex phenomenon. The metastatic process consists of multiple steps, including invasion to the surrounding tissue, intravasation, extravasation, and growth at secondary sites [8]. Given that anchorage-dependent cancer cells must adhere to the extracellular matrix (ECM), the dynamic interplay between cancer cells and ECM profoundly influences the adhesive interactions that facilitate the progression of metastasis [9], [10]. Cell adhesion to the ECM modulates the actin cytoskeleton, which not only provides structural support but also actively participates in cell movement [11]. Understanding the adhesion of cancer cells to the ECM and its connection to actin cytoskeleton organization is crucial for deciphering the molecular mechanism leading to cancer metastasis.

Live-cell imaging, particularly fluorescence-based techniques, can provide a deeper understanding of cellular events and intracellular behavior of biomolecules of interest. Fluorescence-based live-cell imaging techniques have been developed to improve the spatial resolution. For example, structured illumination microscopy (SIM), recognized as the optimal super-resolution method for live-cell imaging, is capable of 100-nm lateral resolution at frame rates up to 11 Hz [12]. The combination of SIM with a higher numerical aperture objective and total internal reflection fluorescence (TIRF-SIM) further improves the lateral resolution (84 nm) [13]. However, acquisition frames are limited to up to 200 image frames owing to intense light illumination (20–100 W/cm2), which causes photobleaching and phototoxicity [14], [15]. Therefore, live-cell imaging is limited to short-term cellular events. As for the axial resolution, TIRF microscopy is recognized as the minimal axial resolution (100–200 nm from the substrate). However, the lateral resolution of TIRF microscopy is limited by the diffraction limit. Therefore, live-cell imaging capable of high spatial (both lateral and axial) resolution and long-term monitoring remains challenging.

We developed a fluorescence imaging method using a plasmonic metasurface composed of self-assembled gold nanoparticles [16], [17], [18]. Here localized surface plasmon resonance (LSPR) of two dimensionally arranged gold-nanoparticles provides high-contrast interfacial images due to the confined light within a region (a few tens of nanometers from the particles) and the enhancement of fluorescence.

A monolayer of oleylamine-capped gold nanoparticles (AuOA) was formed by self-assembly at the air-water interface and then transferred to glass. Closely packed AuOA sheet exhibits enhanced LSPR field compared with isolated AuOA owing to the collective excitation of LSPR in the monolayer. The finite-difference time-domain (FDTD) calculation demonstrates that the penetration depth of the enhanced electric field generated by LSPR is much smaller (a few tens of nanometers) than that of the evanescent field generated by conventional total internal reflection fluorescence (TIRF) microscopy [16]. Live-cell imaging on the AuOA monolayer demonstrates that LSPR-mediated confined light and fluorescence enhancement improve the lateral resolution equivalent to the theoretical limit (a lateral resolution of a few pixels (65 nm per pixel)) and an axial resolution of ∼ 20 nm in experiments (∼15 nm by FDTD calculations) [17]. Additionally, our plasmonic metasurface-based imaging reduces photobleaching owing to the increased emission efficiency via plasmon-exciton coupling. The interface-confined fluorescence and minimal photobleaching of plasmonic metasurface-based imaging enables the observation of a fragile adhesion structure at the initial phase of cell adhesion to a substrate [18]. Our plasmonic metasurface-based fluorescence imaging is simple (compatible with conventional fluorescence microscopy), but has a high lateral resolution equivalent to the theoretical limit and superior axial resolution compared to super-resolution microscopy [17].

In this study, plasmonic metasurface-based live-cell imaging was used to observe actin dynamics during the adhesion of cancer cells. Breast cancer cell lines, non-tumorigenic MCF10A, non-metastatic MCF-7, and highly metastatic MDA-MB-231 cells, were chosen because breast cancer is the most common cancer worldwide [19], and these cell lines are well characterized [20], [21]. We show that our plasmonic metasurface-based imaging system clearly observes actin organization at the cell edge due to confined fluorescence imaging. In addition, we find that breast cancer cell lines exhibit symmetric/asymmetric actin organization at the cell edge during the initial phase of cell adhesion (∼ 1 h after cell seeding).