Stimulated Raman scattering (SRS) microscopy allows label-free imaging of biomolecules based on their intrinsic vibration of specific chemical bonds. In this protocol, the instrumental setup of an integrated SRS and two-photon fluorescence microscope is described to visualize cellular structures in the spinal cord of live mice.
Stimulated Raman scattering (SRS) microscopy enables label-free imaging of the biological tissues in its natural microenvironment based on intrinsic molecular vibration, thus providing a perfect tool for in vivo study of biological processes at subcellular resolution. By integrating two-photon excited fluorescence (TPEF) imaging into the SRS microscope, the dual-modal in vivo imaging of tissues can acquire critical biochemical and biophysical information from multiple perspectives which helps understand the dynamic processes involved in cellular metabolism, immune response and tissue remodeling, etc. In this video protocol, the setup of a TPEF-SRS microscope system as well as the in vivo imaging method of the animal spinal cord is introduced. The spinal cord, as part of the central nervous system, plays a critical role in the communication between the brain and peripheral nervous system. Myelin sheath, abundant in phospholipids, surrounds and insulates the axon to permit saltatory conduction of action potentials. In vivo imaging of myelin sheaths in the spinal cord is important to study the progression of neurodegenerative diseases and spinal cord injury. The protocol also describes animal preparation and in vivo TPEF-SRS imaging methods to acquire high-resolution biological images.
Raman microscopy1,2 is emerging as a powerful label-free method to image biological tissues based on the characteristic frequencies of various chemical bonds in biomolecules. Owing to its non-invasive and well-adaptive imaging capability, Raman microscopy has been widely used for imaging lipid-enriched components in biological tissues like myelin sheath3,4,5, adipocytes6,7, and lipid droplets8,9,10. Stimulated Raman scattering (SRS) signal acquired as stimulated Raman gain (SRG) or stimulated Raman loss (SRL) is background-free, showing perfect spectral resemblance to spontaneous Raman scattering11,12. In addition, SRL and SRG are linearly dependent on the analyte concentration, allowing for quantitative analysis of biochemical components9,11,13. Two-photon excited fluorescence microscopy (TPEF) has been widely used for in vivo biological imaging owing to its inherent optical sectioning capability, deep penetration depth, and low phototoxicity14,15,16. However, the performance of TPEF imaging depends on the characteristics of fluorescent tags, and the number of resolvable colors is limited due to the broadband fluorescence spectra8,17,18,19. Label-free SRS imaging and fluorescence-based TPEF imaging are two complementary imaging modalities, and their combination can provide abundant biophysical and biochemical information of tissues. These two imaging modalities are both based on the nonlinear optical (NLO) processes, which allows simple integration in one microscope system. The combination of the SRS and TPEF imaging, the so-called dual-modal imaging, enables high-dimensional imaging and profiling of cells and tissues, facilitating a comprehensive understanding of complex biological systems. Specifically, picosecond (ps) SRS microscopy can achieve chemical-bond imaging with high spectral resolution compared with femtosecond (fs) SRS technique11, allowing to differentiate multiple biochemical components in biological tissue, especially in the crowded fingerprint region20,21. In addition, compared with another commonly used dual-modal NLO microscope system with integration of coherent anti-Stokes scattering (CARS) microscope, SRS shows superior performance to CARS in terms of spectral and image interpretation as well as detection sensitivity11. The SRS-TPEF microscope has been used as a powerful tool to study various biological systems, such as Caenorhabditis elegans9,22, Xenopus laevis tadpole brain5, mouse brain23,24, spinal cord25,26, peripheral nerve27, and adipose tissue7, etc.
The spinal cord together with the brain makes up the central nervous system (CNS). Visualizing cellular activities in the CNS in vivo under physiological and pathological conditions is critical for understanding the mechanisms of CNS disorders28,29,30 and for developing corresponding therapies31,32,33. Myelin sheath, which wraps and insulates axons for high-speed action potential conduction, plays a significant role in the development of the CNS. Demyelination is thought of as a hallmark in white matter disorders, such as multiple sclerosis34. In addition, after spinal cord injury35, myelin debris can modulate macrophage activation, contributing to chronic inflammation and secondary injury36. Therefore, in vivo imaging of myelin sheath together with neurons and glial cells in living mouse models is of great help to understand the dynamic processes in CNS disorders.
In this protocol, the fundamental setup procedures of a home-built TPEF-SRS microscope are described and the dual-modal in vivo imaging methods for mouse spinal cord are introduced.
All animal procedures performed in this work are conducted according to the guidelines of the Laboratory Animal Facility of the Hong Kong University of Science and Technology (HKUST) and have been approved by the Animal Ethics Committee of HKUST. Safety training for laser handling is required to set up and operate the TPEF-SRS microscope. Always wear laser safety goggles with appropriate wavelength range when dealing with laser.
1. Setup of the TPEF-SRS microscope (for setup schematic see Figure 1)
2. TPEF-SRS microscope system calibration
3. Surgical preparation of mouse for in vivo fluorescence and SRS imaging
4. In vivo TPEF-SRS imaging of mouse spinal cord
In vivo dual-modal imaging of spinal axons as well as myelin sheaths is conducted using the Thy1-YFPH transgenic mice, which express EYFP in dorsal root ganglion afferent neurons (Figure 3). These labeled afferent neurons relay the sensory information from the peripheral nerve to the spinal cord, with the central branch located in the spinal cord dorsal column. With the TPEF-SRS microscope, densely distributed myelin sheath can be clearly visualized using label-free SRS imaging, and sparsely labeled YFP axons can be observed using TPEF imaging. It is revealed by the dual-model imaging that axons are closely wrapped by a thick layer of myelin sheaths (Figure 3C). Nodes of Ranvier (NR), where the axolemma is bare of the myelin sheath, play an essential role in the fast saltatory propagation of action potentials. As can be seen in the TPEF-SRS spinal cord image (Figure 3C), NR show decreased axonal diameter and axolemma directly exposed to the extracellular matrix. It is essential to image the axons together with the surrounding myelin sheaths to confirm the existence and location of NR. Therefore, TPEF-SRS microscopy allows us to observe the dynamic changes of axons and myelin sheaths in the development of spinal cord disorders, which is significant to understand the mechanisms of cellular dynamics.
Figure 1: The schematic diagram of the TPEF-SRS microscope system. The pump and Stokes beams are combined with a dichroic mirror (D1) in the picosecond (ps) laser. The ps beam and femtosecond (fs) beam are collimated and expanded/narrowed by a pair of lenses (L3, L4, and L1, L2, respectively) to match the 3 mm XY-scan galvanometer mirrors. The fs beam is rotated from horizontal to vertical polarization by a half-wave plate (HWP), and then combined with the ps beam by a polarizing beam splitter (PBS). The scanning mirrors and the rear pupil of the objective lens are conjugated by a telecentric scan lens L5 and an infinity-corrected tube lens L6. The laser beam is expanded by the scan and tube lens to fill the back aperture of the 25x objective. For stimulated Raman scattering (SRS) imaging, the backscattered pump beam collected by the objective is reflected by a PBS and directed to a large area (10 mm x 10 mm) silicon photodiode (PD). For two-photon imaging, the two-photon excited fluorescence (TPEF) signal is reflected by a dichroic beam splitter D2 to the photodetection unit. A current photomultiplier (PMT) module is used to detect the TPEF signal. Abbreviations-L1-L10: lenses; OL: objective lens; D1-D3: dichroic mirrors; Fs1, Fs2: filter sets; M: mirrors; OM1, OM2: optical mirrors; P0-P2: alignment plates. Please click here to view a larger version of this figure.
Figure 2: The interface of the delay manager on the OPO control software. The red arrow indicates the checking box for applying the calibrated delay data. Please click here to view a larger version of this figure.
Figure 3: In vivo TPEF-SRS imaging of mouse spinal cord. (A) Schematic diagram of the surgical preparation of the mouse spinal cord and the bright-field image of the spinal cord. (B) Mouse mounting scheme for in vivo imaging of the spinal cord. (C) Maximal z projection images of axons and myelin sheaths in mouse spinal cord. White arrowheads indicate the location of a node of Ranvier. SRS images of myelin are taken at the Raman shift of 2863.5 cm-1. Figure A was created using BioRender (https://biorender.com/). Please click here to view a larger version of this figure.
In this protocol, the basic setup of the TPEF-SRS microscope is described in detail. For SRS imaging, the pump and Stokes beams are temporally and spatially overlapped inside the OPO. However, this overlapping can be disrupted after passing through the microscope system. Therefore, both spatial and temporal optimization of the colocalization of the pump and Stokes beams is necessary and critical to achieving optimal SRS imaging. The temporal delay between the pump and Stokes beam is related to the optical path difference of the two beams, which is determined by the dispersion of optical elements in the microscope system38. When the wavelength of the pump beam is tuned for SRS imaging at different Raman shifts, the optical path length of the pump beam changes accordingly because the refractive index of the optical lenses is dependent on the wavelength38. Therefore, the temporal delay between the pump and Stokes laser pulse changes with the wavelength of the pump beam and thus needs to be calibrated. The OPO is equipped with a software-controlled delay line for temporal synchronization of the pump and Stokes beam. For the same optical setup, the delay data remains stable and only needs to be calibrated once. As a result, the delay data at different wavelengths of the pump beam can be saved at the first measurement for future use. The calibrated delay data can be applied automatically by the OPO control software when the wavelength of the pump beam is changed, which is convenient for SRS imaging at different Raman shifts or hyperspectral SRS imaging. For TPEF-SRS imaging, strict spatial overlapping of the ps and fs beams after the combination is a critical step to avoid any FOV shift between the two imaging models. Firstly, the ps pump beam and fs beam are aligned to make sure they are both on the optical axis of the microscope system, which is critical to avoid any FOV shift when switching the two imaging models. Then, using the pump beam as a reference, the Stokes beam position is adjusted accordingly to achieve strict spatial colocalization. Each alignment procedure requires several trials to reach the optimum.
If the SRS signal significantly decreases, the phase value of the lock-in amplifier and time-delay data should be checked first. Since the lock-in amplifier gets out of phase once the sync signal is disrupted, the phase value requires readjusting every time after its power supply or EOM sync signal is interrupted. The temporal synchronization of the pump and Stokes beam can be quickly checked by slightly adjusting the delay line inside the OPO. If the calibrated time-delay data is far away from the optimal value, a zero scan should be performed to recalibrate the delay offset by clicking the Zero Scan button on the delay manager dialogue. The whole zero scan procedure takes about 10 min. If the SRS signal fails to recover after optimization of the phase and time-delay value, the EOM modulation of the Stokes beam should be checked as described in steps 2.1.3-2.1.4. If the extinction ratio is far less than 10 dB with the observation of small pulse peaks at the off position of the Stokes pulse train, the EOM should be restarted, and the modulation power and phase should be readjusted to achieve maximal modulation depth. Usually, the modulation problem can be solved by resetting the EOM. If not, technical support from the manufacturer should be required.
For in vivo thick tissue imaging, epi-SRS detection mode has to be used. In this protocol, a PBS is used to pass the excitation laser and reflect the back-scattered SRS signal to the detector. The detected SRL signal is dependent on the backscattering of the forward-going pump beam by tissues. The excitation lasers have linear polarization and can fully pass through the PBS, while the backscattered beam has shifted polarization and can thus only be partially reflected by the PBS. Therefore, the current signal collection scheme shows lower efficiency compared to the strategy that directly places an annular photodetector in front of the objective12. Nevertheless, due to the strong scattering of the lipid-rich tissues39, high signal-to-noise ratio SRS images (512 x 512 pixels) of the spinal cord can be acquired with 1-2 s integration time, making this PBS-based collection scheme an appropriate approach for spinal cord imaging. On the other hand, however, the strong tissue scattering limits light penetration depth. For both SRS and TPEF imaging, the imaging depth for the spinal cord is limited to about 50 μm.
The sequential imaging procedure for SRS and TPEF imaging is the major limitation of the current dual-modal imaging method. In the protocol, TPEF and SRS imaging are performed at the same location sequentially with a 1 s interval by switching the motorized flipper automatically. Motion artifacts may cause an imperfect merge of the TPEF and SRS images, which limits the capability of this method for imaging highly dynamic processes or tissues largely affected by the breath and heartbeat of animals. One possible solution is to collect the ps laser-excited two-photon fluorescence simultaneously during SRS imaging9. However, this method is only applicable to the biological structures with strong fluorescence signals, since the ps pulse has a much lower fluorescence excitation efficiency compared to the fs pulse14. Alternatively, the problem can be solved by using an fs-SRS system22,40, where the fs laser source allows simultaneous excitation of the SRS and TPEF signals effectively, at the expense of low spectral resolution of the SRS imaging. Another solution is to use the ps laser-excited fluorescence obtained during SRS imaging as a reference to register the fs fluorescence images. As shown in Figure 3, this registration strategy works well if no significant motion occurs during the SRS and fluorescence imaging.
SRS exhibits unique advantages in biological imaging since it provides chemical information of biomolecules based on its specific label-free contrast mechanism41. Compared with CARS which has also been combined with TPEF for multimodal NLO imaging, SRS showed better spectral and image interpretation capability11. Therefore, it has been widely applied for imaging lipid9,11, protein42,43, DNA44, and bio-orthogonal components containing alkyne (C ≡ C)13,45, carbon−deuterium (C−D)9,46 and oxygen-deuterium (O-D) bonds47,48 in biological tissues. In this protocol, we used a ps laser source for SRS imaging and a fs laser source for TPEF imaging, which combines the advantages of efficient fluorescence excitation and high Raman spectral resolution, allowing effective differentiation of diverse biomolecules42,44. In the spinal cord, complex cell-microenvironment interactions involving glial cells, neurons, and recruited immune cells contribute to the progression of injury49 and diseases50. Combined with various fluorescence and SRS imaging probes, TPEF-SRS microscopy can achieve simultaneous imaging of various cellular structures as well as their distinct biomolecule components, which can significantly facilitate our understanding of the onset and development of spinal cord disorders.
The authors have nothing to disclose.
This work was supported by the Hong Kong Research Grants Council through grants 16103215, 16148816, 16102518, 16102920, T13-607/12R, T13-706/11-1, T13-605/18W, C6002-17GF, C6001-19E, N_HKUST603/19, the Innovation and Technology Commission (ITCPD/17-9), the Area of Excellence Scheme of the University Grants Committee (AoE/M-604/16, AOE/M-09/12), and the Hong Kong University of Science & Technology (HKUST) through grant RPC10EG33.
#2 Forceps | Dumont | 11223-20 | For laminectomy |
10X objective | Nikon | CFI Plan Apo Lambda 10X | |
25X objective | Olympus | XLPLN25XSVMP2 | |
Burn cream | Betadine | ||
Camera | Sony | α6300 | |
Current amplifier | Stanford research | SR570 | |
Current photomultiplier modules | Hamamatsu | H11461-01 | |
D2 665 nm long-pass dichroic mirror | Semrock | FF665-Di02-25×36 | For directing epi-fluorescence signal to the detection module |
D3 700 nm short-pass dichroic mirror | Edmund | 69-206 | For separating SRS from TPEF detection path |
Depilating cream | Veet | ||
FS1 975 nm short-pass filter | Edmund | 86-108 | For blocking stokes beam |
FS1 Bandpass filter | Semrock | FF01-850/310 | For blocking stokes beam |
Fs2 Bandpass filter | Semrock | FF01-525/50 | For selecting YFP signal |
Fs2 Shortpass filter | Semrock | FF01-715/SP-25 | For blocking fs excitation laser beam |
Half-wave plate | Thorlabs | SAHWP05M-1700 | |
High-speed photodetector | MenloSystems | FPD 310-F | For checking Stokes beam modulation |
Iodine | Betadine | ||
IR Scope | FJW | FIND-R-SCOPE Infrared Viewer 2X Kit Model 84499C2X | |
Iris | Thorlabs | CPA1 | |
L1 | Thorlabs | AC254-060-B-ML | |
L10 | Thorlabs | LA4052-A | |
L2 | Thorlabs | LA1422-B | |
L3 | Thorlabs | AC254-050-B | |
L4 | Thorlabs | AC254-060-B-ML | |
L7 | f=100 mm, AB coating | ||
L8 | Thorlabs | LA4874-A | |
L9 | Thorlabs | AC254-035-B-ML | |
Lock-in amplifier | APE | ||
Mirror | Thorlabs | PF10-03-P01 | |
Motorized flipper | Thorlabs | MFF101/M | |
multifunctional acquisition card | National Instrument | PCIe-6363 | |
Oscilloscope | Tektronix | TDS2012C | |
Photodiode | APE | For detecting SRS signal | |
Picosecond laser source | APE | picoEmerald | |
Polarizing beam splitter | Thorlabs | CCM1-PBS252/M | |
Power meter | Newport | 843-R | |
Saline | Braun | ||
Scan lens L5 | Thorlabs | SL50-CLS2 | |
Scanning mirror | Cambridge Technology | 6215H | |
Silicone gel | World Precision Inc. | KWIK-SIL | |
Ti:sapphire fs laser | Coherent | Chameleon Ultra II | |
Tube lens L6 | Thorlabs | TTL200-S8 |