2023 Volume 20 Issue 1 Article ID: e200009
Two-photon excitation laser scanning microscopy (TPLSM) has provided many insights into the life sciences, especially for thick biological specimens, because of its superior penetration depth and less invasiveness owing to the near-infrared wavelength of its excitation laser light. This paper introduces our four kinds of studies to improve TPLSM by utilizing several optical technologies as follows:
(1) A high numerical aperture objective lens significantly deteriorates the focal spot size in deeper regions of specimens. Thus, approaches to adaptive optics were proposed to compensate for optical aberrations for deeper and sharper intravital brain imaging.
(2) TPLSM spatial resolution has been improved by applying super-resolution microscopic techniques. We also developed a compact stimulated emission depletion (STED) TPLSM that utilizes electrically controllable components, transmissive liquid crystal devices, and laser diode-based light sources. The spatial resolution of the developed system was five times higher than conventional TPLSM.
(3) Most TPLSM systems adopt moving mirrors for single-point laser beam scanning, resulting in the temporal resolution caused by the limited physical speed of these mirrors. For high-speed TPLSM imaging, a confocal spinning-disk scanner and newly-developed high-peak-power laser light sources enabled approximately 200 foci scanning.
(4) Several researchers have proposed various volumetric imaging technologies. However, most technologies require large‐scale and complicated optical setups based on deep expertise for microscopic technologies, resulting in a high threshold for biologists. Recently, an easy‐to‐use light‐needle-creating device was proposed for conventional TPLSM systems to achieve one-touch volumetric imaging.
Two-photon excitation laser scanning microscopy system utilizing a near-infrared excitation laser light source is a powerful tool to visualize intravital microstructures for understanding physiological phenomena. Recent technological advances utilizing some optical technologies are remarkable for many research areas. In these several years, we have also applied high-peak-power laser light sources, wavefront modulation technology, super-resolution microscopy, multi-point laser beam scanning methodology, and etc. to the system. Here, we report some of such technological improvement/enhancement studies focusing on the spatiotemporal resolution.
In medical and biological research, fluorescence microscopy has become a familiar and indispensable tool. A targeted fluorophore generally absorbs single-photons with an energy equal to the energy difference between the electronic ground state and the first excited state of the molecule. The phenomenon in which a molecule simultaneously absorbs two photons, having approximately twice the wavelength of the single-photon mentioned above, can also occur. This phenomenon is called as the two-photon excitation process, but the possibility is extremely rare. Two-photon excitation laser scanning microscopy (TPLSM) has several advantages [1,2].
First, the excitation occurrence of the fluorescence molecules is spatially localized at the focus of the excitation laser light, where the photon density is maximum, allowing for optically sectioned fluorescence image acquisition. This spatial localization results from non-linear dependence of the absorption possibility which is proportional to the squared value of the photon density. In contrast, molecular photon absorption in conventional confocal laser scanning microscopy (CLSM) occurs in a wide area outside the excitation laser light focus, which often causes severe problems, such as out-of-focus photobleaching, photodamage, and excitation beam attenuation. Therefore, the localized nature of excitation in TPLSM avoids these problems [3]. Furthermore, the excitation locality in TPLSM does not need a confocal pinhole that is necessary for optical fluorescent image sectioning in CLSM, of which alignment of the position or of the diameter is critical.
Second, the two-photon excitation process of fluorophores, which possess visible spectral properties, is induced by near-infrared (NIR) laser light pulses. Since the biological specimens are relatively transparent in the NIR region [4,5], TPLSM offers superior penetration depth and is less invasive for living specimens. Therefore, TPLSM has become an essential intravital analytical method in the field of biological sciences and medical research and has been extensively developed to use its merits better. Novel optical technologies were applied to improve or enhance the spatiotemporal resolution of TPLSM while maintaining the merits of this method as much as possible. This review article is an extended version of the Japanese article [6].
Optical aberrations in the excitation laser light path perturb ideal focusing, decreasing the fluorescence intensity and degrading the spatial resolution of the microscopy system. A strategy for deeper in vivo brain imaging with maintaining high spatial resolution is optical aberration corrections with adaptive optical technologies, which astronomers have initially developed for the detection optics of astronomical telescopes. In these technologies, fluctuations in the telescope images are compensated by manipulating the optical wavefront of the light from astronomical objects. Adaptive optics in microscopy compensate for optical aberrations by modulating the spatial phase distributions of the 2D wavefront of the excitation laser light beam and/or fluorescent lights. As correction devices, deformable mirrors and liquid-crystal-based spatial light modulators (SLMs) are generally utilized for TPLSM [7].
Recently, two kinds of adaptive optical methodologies for TPLSM were proposed by focusing on the refractive index mismatch at the interface of the living brain tissue. The first method was developed for evaluating the focal volume by measuring tiny fluorescent beads injected into living mouse brains (Figure 1a). With this method, optical aberrations in the deep region of brain tissue were evaluated. We realized that modifications of the beam diameter of the excitation laser light and the refractive index value in the immersion liquid maximized the excitation photon density at the focal position. These two modifications allowed the successful visualization of the fine structures of the hippocampal CA1 neurons and the intracellular calcium dynamics in cortical layer V astrocytes, even with our conventional TPLSM system [8].
(a) Tiny fluorescent bead injection into a living mouse brain for focal spot evaluations. The white arrowheads in the lower 3D reconstructed image indicate single red fluorescent beads with diameter of 1 μm. (b) Adaptive optical (AO) two-photon excitation microscopy for a surface-profiled living thy1-YFP-H mouse brain. The white arrowheads indicate the visualized dendritic spines.
In the second method, we measured 3D coordination of the living mouse brain surface labeled with fluorescent dyes. Based on the measured coordination, a 3D ray-tracing method was applied to calculate a 2D phase-shift distribution of the excitation light beam wavefront. Finally, inversed phase-shift distribution was applied to an SLM installed into the excitation light path, compensating for the optical aberrations caused by the mismatched refractive indices [9]. In vivo two-photon imaging of the secondary motor cortex neurons of anesthetized thy1-YFP-H mice brain successfully visualized individual dendritic spines at a depth of 500 μm (Figure 1b) based on this methodology using the assumed refractive index value of the brain.
The super-resolution microscopic technologies, STED microscopy [10] were applied to improve TPLSM spatial resolution. Since the basic optical setup of the STED microscopy system is based on laser scanning fluorescence microscopy, this can be merged with conventional TPLSM systems. Therefore, several attempts have progressed over the past decade [11–18].
We developed two-photon excitation STED microscopy of which key components were electrically controllable components for easy and efficient imaging (Figure 2a) [19–21]. One was a set of transmissive liquid-crystal devices (tLCDs), a type of SLM, that converts the STED beam into an optical vortex focusing as a doughnut shape, and compensated for the chromatic aberration of the objective lens. Unlike reflective SLMs, tLCDs were able to be directly inserted in front of the objective lens without adding a new optical path, enabling a compact optical setup. Next was a pulsed light source system based on a semiconductor laser diode, which was employed to generate both two-photon excitation and STED light pulses [22,23]. In order to achieve super-resolution imaging with less photodamage to specimens, STED light pulses required low average power and stable temporal synchronization with excitation light pulses. In our setup, the time delays of both optical pulse generations were electrically controlled independently, thereby achieving picosecond-precision pulse synchronization. Combining these components in two-photon STED microscopy improved the spatial resolution to 70 nm, approximately five times higher than in conventional TPLSM (Figure 2b). After screening suitable fluorescent dyes or proteins for biological applications, we recognized the sub-100 nm spatial resolution of the microtubules in a fixed mammalian cell (Figure 2c). Notably, there was no severe photobleaching, implying that our two-photon STED microscopy will facilitate long-term living cell imaging under physiological conditions. More recently, with more modifications, this system has been preliminarily applied to living cells and thick tissue specimens, of which images will be shown shortly.
(a) Two-photon excitation stimulated emission depletion (TP-STED) microscopy utilizing electrically controllable optical components. DM: dichroic mirror; λ/2: halfwave plate; NDD: non-descanned detector. (b) TPLSM and TP-STED microscopy comparison images of a Nile Red labelled fluorescent bead with a diameter of 20 nm. (c) TPLSM and TP-STED microscopy comparison images of AlexaFluor546 labelled microtubule networks in the fixed mammalian cell.
Most TPLSM systems utilize a single-point laser beam scanning method using moving mirrors, and their temporal resolution primarily depends on the speed of the physical movement of these mirrors. In contrast, multi-point laser beam scanning methods, such as a spinning-disk confocal scanning unit (Figure 3a), have improved the TPLSM system (TPLSM-SD) temporal resolution in the early 21st century [24]. However, the insufficient energy of conventional mode-locked titanium-sapphire laser light sources restricted the field of view (FOV) to a narrow region [24,25].
(a) Two-photon excitation spinning-disk confocal microscopy. (b) 5D (xyzt-λ) mitotic progression visualization of 3-color labeled tobacco BY-2 cells. (c) Polarization-resolved fluorescence imaging of mouse pancreatic acini, of which plasma membrane was labelled with FM4-64, and second harmonic generation imaging of collagen fibers in a mouse skin.
Therefore, high-peak-power laser light sources (neodymium-based 918 nm and ytterbium-based 1042 nm) were introduced to extend the FOV [26,27]. The repetition rates of these high-peak-power laser light sources were less than 10 MHz, disabling to be utilized for high-speed scanning in single-point TPLSM due to the short pixel dwell time. In contrast, the spinning-disk unit scanned a specimen with approximately several hundred excitation beams. Therefore, the measurement conditions of this study were sufficient to acquire a video-rate image or faster. Moreover, the pinhole-array disk in the detection optics provides the TPLSM-SD system confocality, resulting in superior axial spatial resolution compared to the conventional TPLSM system. By utilizing the developed TPLSM-SD system, several biological phenomena have been visualized [28–32]. As a representative example, Figure 3b showed time-lapse xyz images of three kinds of organelle in a tobacco BY-2 cell undergoing the mitotic progression [32]. Since mitotic control is susceptible to photodamage, intensive excitation laser light irradiations often cause mitotic progression blockage, especially for detailed multidimensional measurements. This study quantitatively evaluated phototoxicities by comparing conventional CLSM measurements. As a result, the spatial locality of two-photon excitation was attributed to decreased phototoxities. Another example was shown in Figure 3c, which represents polarization-resolved imaging based on the TPLSM-SD system [29]. Since the TPLSM-SD system adopted a 2D detector, the system could be easily combined with image-splitting optical units, which have been developed by several companies and are commercially available. Without complicated and expensive detection optics modulation, placing a polarizing beam splitter cube into the optical unit enables intravital polarization-resolved imaging of fluorescently-labeled plasma membrane in mouse pancreatic acini and second-harmonic generation signals from collagen fibers in mouse skin.
Visualizations of the 3D microstructures of the specimens generally require z-scanning methods, such as a focus drive motor or a piezo-actuator to change the relative positions between the objective lens and the specimens. Therefore, the temporal resolution for volumetric imaging largely depends on z-scanning mechanical speed, which generally requires at least a few seconds. However, 3D visualization of biological processes in living specimens, such as calcium transients or tracking subcellular structure dynamics, requires a considerably higher temporal resolution. To increase the temporal resolution, several researchers have proposed cutting-edge volumetric imaging technologies [33]. Among them, a Bessel beam-based TPLSM engineers the point spread function (PSF) to increase the effective depth of field (DOF) [34–38], enabling projection image acquisition by single scanning with the excitation within a needle-shaped region. To realize single-scan projection imaging by a Bessel beam, the spatial distribution of the excitation light beam should be converted to an annular shape at the pupil plane of the objective lens or its optically relayed positions. However, these ideal positions were typically inaccessible to commercially available TPLSM systems, indicating that light-needle scanning in these TPLSM systems was almost impossible.
Recently, we thus proposed a method for realizing volumetric imaging with an extended DOF in a TPLSM by developing an easy-to-use, light-needle-creating device that can be built into conventional microscopes [39]. This device consisted of a concave axicon and a plano-convex lens. We directly inserted it into one position of the filter turret of an upright microscope of the TPLSM system (Figure 4). The concave axicon converted an incident Gaussian beam into an annularly spreading beam owing to its refraction by the conical surface. The converted beam then formed a thin annular-shaped pattern at the focus of the convex lens built into the stacked device. By placing the plano-convex lens, the resultant annular-shaped beam was focused by an objective lens to form a needle-shaped two-photon excitation spot. Fluorescence emission from fluorophores existing in the focal volume was collected by the same objective lens and reflected by a dichroic mirror toward a non-descanned detector before the light-needle-creating device. As a result, the device inserted in the turret generated an axially elongated PSF extended by a factor greater than 70 compared to that of a conventional TPLSM. The developed microscopy system successfully demonstrated single-scan projection imaging of thick biological specimens, resulting that 3D locations of visualized microstructures were resolved by referring to the Gaussian z-stack image. Although the proposed methodology has some practical limitations, such as the field curvature and depth-dependent magnification, in exchange for ease of use, the device potentially reduced the cost and complexity of the Bessel beam scanning-based volumetric technique for broader applications.
(a) Single-scan volumetric imaging by one-touch installable light-needle-creating devise for conventional two-photon excitation microscopy system. NDD: non-descanned detector; MIP: max intensity projection. (b) Comparison between z-stacked Gaussian beam scanned image and Bessel beam scanned image of a fixed thy1-YFP-H mouse brain slice.
Technological developments in multi-photon microscopy are progressing worldwide, resulting in one of the essential methods to visualize and analyze biological phenomena. Especially in recent years, fluorescence microscopy utilizing the three-photon excitation processes has been developed and is becoming a practical methodology [40]. These cutting-edge imaging technologies also have contributions from numerous essential technologies that are deeply rooted in spectroscopic research, photochemistry, applied physics, and information science. Matching the needs of biology will give birth to many technological innovations. Conversely, as long as the technical development is oriented toward biological research, developers must refrain from putting the cart before the horse, in which the observation disturbs further investigations of the targeted physiological phenomena. It is crucial for further development in the future to always bear in mind the inherent practicality of fluorescence microscopy, which enables us to go inside life with minimal invasiveness.
The authors declare no conflicts of interest.
K.O. and H.I. performed experimental and analytical studies with graduate students. K.O., H.I. and T.N. developed the concept of introduced methods. K.O., H.I. and T.N. wrote the manuscript.
The data of microscopy images shown in Figure 3b is stored in the SSDB:OMERO repository (Systems Science of Biological Dynamics database: https://ssbd.riken.jp) with the project ID 767. Other data of microscopy images in this study is available from the corresponding author on reasonable request.
We would like to thank Dr. S. Sato, Dr. Y. Kozawa, and Dr. H. Yokoyama of Tohoku University for their helpful advices regarding optical setup and for lending their technical supports; Dr. N. Hashimoto, Mr. M. Kurihara, and Dr. A. Tanabe of Citizen Watch Co., Ltd. for kindly providing the tLCDs; Dr. T. Murata and Dr. M. Hasebe of National Institute for Basic Biology for providing multi-color labeled tobacco BY-2 cells; and graduate students of Hokkaido University and the Graduate University for Advanced Studies (SOKENDAI), Mr. R. Kitamura, Dr. K. Yamaguchi, Ms. Y. Yamanaka, Ms. A. Goto, Mr. T. Kamada, Mr. K. Nakata, and Dr. C.-P. Chang, for their immense efforts in these studies. We are also grateful for the advices provided by Dr. R. Enoki and Dr. M. Tsutsumi of the Exploratory Research Center for Life and Living Systems. This study was supported by the Brain Mapping by Integrated Neurotechnologies for Disease Studies (JP19dm0207078) from the Japan Agency for Medical Research and Development (AMED); by the Core Research for Evolutional Science and Technology (CREST) (JPMJCR20E4) from Japan Science and Technology Agency (JST); and by the KAKENHI (16H06280/22H04926 “Advanced Bioimaging Support,” 18K06591, 18K14659, 20H05669, 21K19346, and 22H02756) from the Japan Society for the Promotion of Sciences (JSPS) in the Ministry of Education, Culture, Sports, Science and Technology (MEXT).
