Multiphoton microscopy, invented by Denk, Strickler and Webb in 1990, has changed how we visualize cells by providing high-resolution, non-invasive imaging capability deep within intact biological tissues. In vivo, cellular resolution fluorescence imaging is now routinely achievable at greater than 1 mm depth in mouse brains. In this talk, I will present research on pushing the depth limit of long wavelength 3-photon microscopy. I will reflect on my journey in multiphoton imaging and the enormous impact Watt Webb has had on me in science and in life.

The molecular mechanisms that define cellular functions can be elucidated by live-cell hyperspectral fluorescence microscopy, which permits simultaneous measurements of multiple dynamic processes with excellent signal-to-noise ratios. We have developed a novel five-dimensional (x,y,z,t,λ) imaging system by combining dual-view Selective-Plane Illumination Microscopy (diSPIM) for isotropic diffraction-limited imaging and image mapping spectroscopy (IMS) for single snapshot hyperspectral imaging of multiple fluorophores. We will demonstrate the properties of this instrument and show its application to imaging islet of Langerhans, a micro-organ that contributes to blood glucose homeostasis through the regulated secretion of glucagon (from α-cells) and insulin (from β-cells).

Cellular studies using model in vitro systems limit environmental complexities that may obscure input-outcome relationships. To address the need for more relevant model systems, we have developed a platform for interfacing biomaterials with cultured cells with high spatiotemporal control. Pulsed near-infrared light is used to create protein-based cellular landscapes in a direct-write, multiphoton photo-crosslinking process patterned using a dynamic mask. High-resolution 3D environments are created either in advance of cellular application, or in the presence of viable cells, yielding dynamically controllable enclosures and surfaces for organizing cellular communities that more accurately reproduce mechanical, chemical, and convective properties of native environments.

Cell surface receptors distinguish specific stimuli from biological noise, and underlying mechanisms depend on the organization of plasma membrane components. We exploited the statistical robustness of multiplexed Imaging Fluorescence Correlation Spectroscopy (ImFCS), to evaluate subtle diffusion differences of key components in the early stage of mast cell signaling by measuring their diffusion in live cells: FcεRI receptor for immunoglobulin E (IgE), tyrosine kinase Lyn anchored to the membrane inner leaflet, and transmembrane phosphatase PTPα. Our results confirm that the nanoscale distribution of signaling proteins is coordinately rearranged after the receptor is stimulated, such that changes in relatively weak interactions serve to transmit the signal across the membrane.

Multiphoton microscopy has benefited from many advances since its invention in the Webb lab. I will highlight some of the advances I had the privilege of working on over the years. A key advantage of multiphoton microscopy is that it enables high-contrast imaging in thick tissue. I will discuss some recent directions in pushing the limits of this advantage, particularly related to deep imaging at high speeds.

Multiphoton microscopy is used in diverse branches of biology. While its main applications are related to neuroscience, cancer and stem cell research take advantage of this imaging technique. In addition, the push to use multi-photon imaging for clinical and even diagnostic applications is gaining momentum both on the research and commercial front. This broad range of applications requires diverse types of femtosecond laser sources - compact, dedicated and portable on one side, and flexible and very high-performance on the opposite end. In this talk we will discuss recent developments in laser technology addressing these various requirements.

Multiphoton microscopy is the gold standard for deep tissue fluorescence imaging. Long wavelengths enable hundreds of microns deep penetration of excitation light, but the emission fluorescence at shorter wavelengths encounters scattering before detection. While not being an issue for point scanning geometries, for wide-field geometries emission light scattering degrades the image quality. In this work, we use temporally focused pattered excitations to spatially encode image information before emission light scattering. Upon detection, images are reconstructed computationally by solving a linear inverse problem. We further improve our results by learning inverse solvers and optimal patterns through physics-based deep learning.

There is a continuing need to develop preclinical molecular imaging modalities to guide the development and optimization of targeted therapies in oncology. We have established Macroscopic Fluorescence Lifetime Imaging (MFLI) associated with Forster Resonance Energy Transfer (FRET) to report quantitatively on antibody-target engagement in live intact animals at the organ level. Here, we use FLI FRET imaging to quantify the binding of near infrared labeled antibody drugs to breast cancer cells or tumor xenografts. In combination with antibody engineering, this approach provides a robust analytical tool to optimize the binding and uptake of antibody-target binding during tumor progression and treatment.

Multiphoton and confocal laser scanning fluorescence lifetime imaging microscopy (FLIM) are often limited by slow acquisition due to the dead time of photon-counting and time-tagging analog electronics. Here, we present a solution for faster imaging with lower dead time by directly digitizing the amplified detector output and computationally determining photon counts via GPU-accelerated processing using our custom Single- and multi-photon PEak Event Detection (SPEED) algorithm. The algorithm uses thresholded local maxima detection to temporally localize photon counts in the digitized data, enabling sub-ns dead time between consecutive photon counts.

We present a new architecture of Time-Correlated Single Photon Counting (TCSPC) and time tagging electronics with multiple synchronized input channels. Its unique data interface enables fast data transfer and rapid on-the-fly analyses via external FPGA boards beyond the throughput-limitations of widely established data interfaces. Apart from showing design features and benchmark results of the instrument as such, we present results from fluorescence lifetime imaging and a development snapshot of other high throughput applications.

Human pluripotent stem cell (hPSC)-derived retinal cell culture holds great promise for treatment of retinal diseases that cause blindness. However, high-throughput non-invasive functional assays are needed for rapid optimization and quality control of these stem cell culture systems. Here we use simultaneous two- and three-photon excited fluorescence lifetime imaging microscopy (FLIM) to characterize the autofluorescent signature of light response in photoreceptor cells that originates from the fluorescent vitamin A compounds (i.e., retinoids) of the visual cycle. This multiphoton microscopy technique resolves the dynamics of visual pigment photobleaching and recycling following light exposure, including conversion of different retinoids to one another.

Multiphoton fluorescence lifetime imaging of the metabolic coenzymes reduced nicotinamide adenine dinucleotide (NADH) and flavin adenine dinucleotide (FAD) allows quantification of cellular metabolism. Due to the link between cellular metabolism and cell function, autofluorescence lifetime imaging provides many features for identification of cells with different phenotypes. Segmentation of multiphoton fluorescence lifetime images allows analysis of data at a single-cell level and quantification of cellular heterogeneity. In this study, Gaussian distribution modeling and machine learning classification algorithms are used for the identification of rare cells within autofluorescence lifetime image data.

In this work, we investigate the changes of microviscosity in cancer cells exposed to chemotherapy with cisplatin and oxaliplatin. We analyzed the plasma membrane viscosity using fluorescent molecular rotor and fluorescence lifetime imaging. Our results in vitro show a significant increase in viscosity after treatment, which correlated with membrane lipid composition. The obtained data indicate the involvement of plasma membrane in the mechanisms of the platinum drug action.

The human skin dermis is vascularized soft tissue containing highly organized extracellular matrix (ECM) architecture. Macrophages mediate wound healing within the skin but technologies for detection of macrophage function simultaneously with ECM properties are limited. In this study, we use multiphoton microscopy to evaluate the metabolic function of macrophages and quantify collagen alignment. Multiphoton microscopy is label-free and non-damaging allowing time-course studies of turbid samples and in vivo measurements. The fluorescence lifetime imaging technique detects the time fluorophores remain in an excited state before relaxing to the ground state. Reduced nicotinamide dinucleotide (NADH) and oxidized flavin adenine dinucleotide (FAD) are auto-fluorescence coenzymes that participate in various metabolic pathways including glycolysis, oxidative phosphorylation, fatty acid synthesis, and fatty acid oxidation, and label-free fluorescence lifetime imaging is a functional method to track cellular metabolism. In this paper, macrophages were seeded on anisotropic and isotropic ECM. Fluorescence lifetime images of macrophages and second harmonic generation (SHG) images of collagen in the ECM were obtained with a two-photon fluorescence lifetime microscope every two days for one week. Significant changes in macrophage size and fluorescence lifetime features were observed with time. The free fraction of NAD(P)H and bound fraction of FAD are both higher for macrophages on random ECM compared with aligned collagen. In summary, multiphoton microscopy provides a label-free method to track metabolic variation of macrophages concurrently with imaging of the ECM.

We use SHG microscopy to study collagen alterations in idiopathic pulmonary fibrosis (IPF) in human tissues and in vitro models. Texture analysis successfully classified the fiber morphology in normal and IPF tissues at near 100% accuracy. SHG polarization measurement indicated reduced chirality in the collagen triple helix in IPF in both tissues and spheroid models. Using SHG directional measurements (F/B), we found the collagen fibril and macromolecular structures are altered in IPF, consistent with decreased organization. We are developing a rigorous phasematching approach to determine the collagen fibril size, packing, and polarity based on these measurements.

We investigated the role of collagen fiber morphology and matrix stiffness on cancer cell migration using multiphoton polymerized scaffolds based on Second-Harmonic Generation (SHG) images of ovarian tumors. We replicated the respective collagen fiber architectures and constructed models representing normal stroma, high-risk tissue, benign tumors, and high-grade tumors. The primary finding is that cell–matrix interactions such as motility, cell spreading, f-actin alignment, focal adhesion, and cadherin expression are mainly determined by the fiber morphology to a larger extent than the initial cell type. Moreover, increased matrix stiffness resulted in enhanced motility and cytoskeletal alignment, consistent with a durotactic mechanism.

Combining coherent anti-Stokes Raman scattering(CARS) with the surface-enhancement effects afforded by plasmonic resonances can probe molecular vibrations down to the single-molecule limit. Recent advances have been encouraging, yet, there are difficulties associated with the reproducibility of these measurements. Here, we present controllable and repeatable wide-field surface-enhanced CARS (SECARS) measurements at metal film-metal particle and metal film-dielectric particle nanojunctions. We demonstrate CARS signal at metal-metal junctions using the lowest average power densities to date. We also show reproducible SECARS measurements at metal-dielectric heterojunctions, opening the possibility of SECARS sensors not constrained by lossy and fragile metallic antenna systems.

Understanding complex biological systems requires simultaneous characterization of a large number of interacting components in their native 3D environment. However, fluorescence methods are usually hindered by the fundamental color barrier. Here we will introduce electronic pre-resonance stimulated Raman scattering (epr-SRS) microscopy, an emerging method of ultrasensitive vibrational imaging with high potential of multiplexing. To generalize epr-SRS to large-scale volumetric imaging, we further integrated it with tissue clearing technology and achieved simultaneous visualization of >10 protein targets over millimeter thickness of brain tissues. Overall, super-multiplexed vibrational imaging is a promising tool to provide a complete picture of tissue biology.

We applied multiphoton fluorescence, SHG, and SRS to investigate In situ and chemical-specific identification of lipid and mitochondrial composition is crucial in analyzing the impacted subcellular components resulting in lipid accumulation, and differentiate mitochondrial associated bimolecular identification and related lipids’ spatial distribution in skeletal muscle cells between athletes and type 2 diabetes patients, to reveal differences in composition related to dysregulated metabolic pathways. Furthermore, quantitative SRS analysis revealed markers of irregular molecules for impacted mitochondrial function corresponding to decreased lipid metabolism seen in type 2 diabetes.

We demonstrate an approach for producing spatially resolved maps of macromolecular mobility compatible with multi-photon excitation in fluorescence recovery after photobleaching (FRAP). Periodically structured illumination during photobleaching allows distribution of the photobleach power over the entire field of view to reduce local heating effects, but recovers highly localized peaks in the spatial Fourier transform domain. Recentering the FT peaks and inverse Fourier transformation yields images scaling with the amplitude of the photobleach modulation at each spatial location. From the temporal decays, up to 250,000 independent diffusion coefficients can be recovered for each position in the field of view.

We propose the use of visible-wavelength two-photon excitation (v2PE) for activation of reversibly photo-switchable fluorescent proteins (RSFPs) and successive confocal detection to achieve super-resolution imaging. In this method, three photons interact with the sample molecules in total, which provides imaging properties equivalent to using third-order nonlinearity in fluorescence response. Because this technique uses visible light, it can achieve higher spatial resolution than confocal microscopy. In this study, we performed experimental investigations to confirm the activation of negative RSFPs by v2PE and demonstrated super-resolution imaging of live cells.

We have realized video-rate three-photon fluorescence microscopy based on a simple 1320-nm femtosecond source, which is driven by a 24-MHz Cr:forsterite oscillator and fiber-optic nonlinear conversion. The development aims to find a robust pumping source for deep-tissue GFP imaging. We optimized the laser oscillator to deliver 40-nJ output pulse energy at around 1260 nm, and the matching of the laser spectrum and GFP excitation can be realized by precisely control the interplay between self-phase modulation and fiber dispersion. We have obtained clear video-rate three-photon imaging from fluorescent beads, and further optimization leading to deep-tissue GFP imaging will be discussed.