Biological tissue transparency combined with light-sheet fluorescent microscopy has become an important method for studying the neural structure of biological tissue, and methods for clarifying biological tissue have advanced greatly.1–3 Biological tissue transparency is mainly attained via denaturation of protein or removal of lipids of the cell membrane, to eliminate or reduce the refractive index or the scattering index of the biological tissue, so that the deeper organizational structure of nerves can be observed.4,5 The recently developed clarifying methods can be divided into three categories: aqueous-based clearing, hydrogel embedding combined with electrophoresis clearing, and solvent-based clearing. Aqueous-based clearing of biological tissues tends to cause tissue expansion and requires an extended period of time, and some denaturing reagents, such as urea or sodium dodecyl sulfate, could cause the tissue to become antigen-insensitive.6–13 Hydrogel embedding combined with electrophoresis clearing is a complex method that requires gel embedding and electrophoresis equipment, and the reproducibility of experimental results is poor.14–16 Solvent-based reagents virtually quench endogenous fluorescent proteins.17–20 Although the ultimate three-dimensional imaging of solvent cleared organs technique can clarify biological tissues expressing fluorescent proteins, the fluorescence intensity gradually diminishes after clearing.21 To resolve these problems, the fluorescent signal requires long-term preservation and antibleaching treatment after resin embedding; this method involves first using solvent-based clearing of the biological tissue, after which it is embedded in epoxy resin.22
We developed an acrylic resin with a high refractive index resin (termed HRAR) that clears biological tissue directly after embedding. The refractive index is increased when the liquid monomer polymerizes into solid resin, which matches the refractive index of the dehydrated biological tissue perfectly, to achieve transparency. This method can clear large samples directly without requiring special pretreatment, and can maintain the endogenous fluorescent signal well. In this study, we investigated the effect of differences in the pH value of monomers on green fluorescent protein (GFP) fluorescence,23 and a transparent GFP mouse brain block imaged via light-sheet fluorescence microscopy.
Monomer 2-phenoxyethyl acrylate and the crosslinking agent ethylene glycol dimethacrylate (mass ratio of ) were polymerized in the presence of 0.2% initiator 2,2′-azobis(2,4-dimethyl) valeronitrile to produce HRAR resin (Fig. 1). The main effect of 2-phenoxyethyl acrylate is to increase the refractive index after polymerization, and the crosslinking agent triethylene glycol dimethacrylate causes the resin to form a crosslinked network structure that improves the hardness of the resin. The results (Fig. 2) showed that the light transmittance of an HM20 resin-embedded mouse brain was about 7% in the spectral range of 400 to 1100 nm, while the transmittance of an HRAR resin-embedded mouse brain was about 70% in this spectral range.
Eight-week-old mice were perfused according to the recommended protocol for the perfusion of mice.24 Animal care and use were in accordance with the guidelines of the Administration Committee of Affairs Concerning Experimental Animals in Hubei Province of China. The protocol was approved by the Committee on the Ethics of Animal Experiments of the Huazhong University of Science and Technology (permit number: 00027340). All efforts were made to minimize the suffering of the animals. The tissue was sectioned into coronal slices using a vibrating microtome (Leica, VT1000 S), and the slices were stored in 0.01 M PBS solution at 4°C in the dark.
The brain slices were dehydrated according to the following procedure: 75% tetrahydrofuran (THF) for 3 min, 100% THF for 3 min, and 100% THF for 4 min. The brain slices were then placed in the resin mixture twice for 5 min each time. Two drops of resin mixture were then added to the slides, which were subsequently coverslipped. The brain slices on the slides were polymerized in an oven at 45°C for 8 h. This experiment was carried out to verify the endogenous fluorescence intensity before and after embedding in the HRAR resin. First, brain slices were placed on a slide and coverslipped. Confocal fluorescence microscopy (Zeiss 780) was then used to image the region of interest. The imaged slices were embedded according to the above procedure, and confocal fluorescence microscopy (Zeiss 780) was again used to image the same region using the same parameters. To quantitatively evaluate the endogenous fluorescence intensity of the tissue before and after embedding in the HRAR resin, we used the fluorescence intensity of the soma as an indicator. First, the imaging data were imported into Image J software. Within this software, the rectangular-selection tool was activated, a circular area on the soma was selected, and the histogram tool was then used to measure the average gray value; the average gray value of the soma before embedding was denoted as , while the average gray value of the soma after embedding was denoted as . The fluorescence retention rate after embedding was calculated as follows: . For each group of experiments, we selected 10 somata to evaluate the average gray values.
2-phenoxyethyl acrylate and ethylene glycol dimethacrylate were filtered through alkaline alumina (), respectively, solution a was prepared by mixing 2-phenoxyethyl acrylate and ethylene glycol dimethacrylate in a mass ratio of , then 0.2% initiator 2,2′-azobis(2,4-dimethyl) valeronitrile was added, at last the mixture was stirred for 30 min; solution b was prepared by mixing 2-phenoxyethyl acrylate and ethylene glycol dimethacrylate in a mass ratio of , 10% tetrasodium ethylenediaminetetraacetate dehydrate () was then added and the solution was shaken for 6 h. The mixture was filtered and then 0.2% initiator 2,2′-azobis(2,4-dimethyl) valeronitrile was added, at last the mixture was stirred for 30 min. Solution c was prepared by mixing 2-phenoxyethyl acrylate and ethylene glycol dimethacrylate in a mass ratio of , 10% tetrasodium ethylenediaminetetraacetate dehydrate () was then added and the solution was shaken for 6 h. The mixture was filtered and then 0.2% initiator 2,2′-azobis(2,4-dimethyl) valeronitrile and 0.5‰ 2,2′-diaminodiethylamine were added, at last the mixture was stirred for 30 min. The three solutions that were measured with a pH meter (OHAUS S-5000 pH meter, Parsippany, New Jersey) were calculated the average pH value after multiple measurements. The results (Fig. 3) show that the fluorescence retention ratio of endogenous GFP signal was 46% when the pH value of the resin monomer was 4.81, 162% when the pH value of the resin monomer was 6.27, and 210% when the pH value of the resin monomer was 9.62. This indicates that the pH value of the resin monomer affected the fluorescence signal intensity markedly, and that an alkaline environment is beneficial to the maintenance of the endogenous GFP signal.
The postfixed mouse brain block was rinsed three times with 0.01 M PBS solution for 8 h each at 4°C in the dark, after which it was dehydrated in 50% THF for 2 h, 75% THF for 2 h, 95% THF for 2 h, 100% THF for 3 h, and 100% THF for 3 h at 4°C. The dehydrated brain block was then embedded in HRAR resin as follows: HRAR resin for 2 h, HRAR resin for 24 h, and HRAR resin for 48 h at 4°C in the dark. Next, the permeated mouse brain block was placed in a capsule that was filled with HRAR resin. The curing procedure was as follows: 32°C for 8 h, 38°C for 8 h, 45°C for 8 h, and 50°C for 4 h in vacuum and dark conditions. The results (Fig. 4) from light-sheet fluorescence microscopy imaging showed that the fluorescence signal and neuronal fiber structure were well preserved with the HRAR resin (solution c), and that neuronal fibers were not distorted by this embedding process. Thus, the polymerization rate was uniform, and this method could be applied for clearing large samples.
In this paper, we applied the principle of refractive index matching to clear biological tissue, and found that the refractive index increased by about 0.05 after the acrylic monomer polymerized. We then developed an acrylate resin with a high refractive index that could clear dehydrated biological tissue directly. The average refractive index of the resin monomer is 1.512, and the biological tissue was translucent after penetration by the resin monomer. The average refractive index of the polymerized resin is 1.558, and the refractive index of the dehydrated biological tissue is 1.55 to 1.56,1 the refractive index of the acrylate resin, therefore, matches that of the dehydrated biological tissue, which results in direct clearing of the biological tissue after polymerization. The refractive index of a traditional commercial acrylic monomer is about 1.40 to 1.45, and the refractive index of the solid commercial resin is about 1.50. Therefore, biological tissue embedded in such resin cannot be transparent, as the refractive index of the commercial resin is mismatched to that of the dehydrated biological tissue.
In contrast with traditional commercial acrylate resin, the HRAR resin greatly improved the ability of light to penetrate biological tissue. The light transmittance in the spectral range of 400 to 1100 nm does not change markedly, which ensured that standard light wavelengths could penetrate the embedded biological tissue, greatly reducing attenuation of the excitation and emission light.
We also investigated the fluorescence retention by adjusting the pH value of the resin monomer, and found that an alkaline environment is beneficial to maintaining endogenous GFP signals, while an acidic environment could quench endogenous GFP signals.23 Thus, the pH value of the resin monomer can affect endogenous GFP signal intensity markedly. In our study, solution c was the best for maintaining the GFP signal. Light-sheet microscopy imaging of brain block showed that the cell bodies and nerve fibers were maintained intact, indicating that this method is also suitable for large sample clearing and light-sheet fluorescent microscopy imaging.
The authors have no relevant financial interests in this article and no potential conflicts of interest to disclose.
We thank Jianyi Xu, Fang Fang Yin, Yadong Gang, Pei Li, Ke Bai, Lin Li, and Yuanlei Yue for their helpful comments. This work was supported by the NSFC Project Nos. 81327802 and 61421064.