A novel paper MAP method for rapid high resolution histological analysis

Three-dimensional visualization of cellular and subcellular-structures in histological-tissues is essential for understanding the complexities of biological-phenomena, especially with regards structural and spatial relationships and pathologlical-diagnosis. Recent advancements in tissue-clearing technology, such as Magnified Analysis of Proteome (MAP), have significantly improved our ability to study biological-structures in three-dimensional space; however, their wide applicability to a variety of tissues is limited by long incubation-times and a need for advanced imaging-systems that are not readily available in most-laboratories. Here, we present optimized MAP-based method for paper-thin samples, Paper-MAP, which allow for rapid clearing and subsequent imaging of three-dimensional sections derived from various tissues using conventional confocal-microscopy. Paper-MAP successfully clear tissues within 1-day, compared to the original-MAP, without significant differences in achieved optical-transparency. As a proof-of-concept, we investigated the vasculature and neuronal-networks of a variety of human and rodent tissues processed via Paper-MAP, in both healthy and diseased contexts, including Alzheimer’s disease and glioma.


Visualization of hippocampal vasculature in Alzheimer's disease via Paper-MAP.
To demonstrate proof-of-concept, we processed 100 µm-thick brain slices derived from the 5xFAD mouse model of Alzheimer's disease (AD) according to our Paper-MAP protocol. AD brain slices were easily cleared and expanded by Paper-MAP within 1 day, after which we performed lectin immunostaining to investigate hippocampal angiogenesis. Vascular defects such as profuse thinning and increased microvasculature were apparent at 3 months of age, and by 9 months of age the AD hippocampus showed multiple hemolytic plaques (Fig. 4a). We were also able to quantify blood vessel length and volume, which showed significant differences between 5xFAD and wild-type brains (Fig. 4b-d). These results support the feasibility and efficacy of Paper-MAP as a tool for threedimensional visualization of cellular and subcellular structures in intact tissue.

Investigation of mouse spinal cord injury scar tissue via Paper-MAP.
To further demonstrate the applicability of Paper-MAP, we applied this technique to two mouse models of spinal cord injury (SCI). Briefly, we generated two SCI models, involving spinal cord hemisection and deletion, respectively, and screened mice for poor performance on behavioral tests suggestive of injury (see Supplementary Figs. S1 and S2 online). We then harvested spinal cord tissue from injured mice and generated 100 µm-thick sections for processing via Paper-MAP (Fig. 5a). Immunostaining of hemisection-induced SCI for lectin revealed increased cell cohesion in the scar region, which was also observed upon immunostaining for neurofilament (NF-H, red) and astrocytic glial fibrillary acidic protein (GFAP, red) ( Fig. 5b- www.nature.com/scientificreports/   immunostaining in the hippocampus of mouse brain sections processed via Paper-MAP, imaged with confocal microscopy (4 × 5 panels, horizontal × vertical tiles). Each image (enlarged; i, ii, and iii) was taken with the same 63 × objective lens and z-stacked (range: 100-μm) for comparison. Scale bars (green, 1000-μm; gray, 100-μm).
(b) Lectin and neurofilament (NF) immunostaining in the hippocampus and cortex of mouse brain sections processed via Paper-MAP. Each image of before (left; unexpanded mouse brain sections of 4% PFA fixed) was taken with the same 40 × objective lens and single z-images (5 × 5 panels, horizontal × vertical tiles) for comparison. Enlarged images are of each regions (yellow box). Images of after Paper-MAP (right) was taken with the same 40 × and 63 × objective lens and z-stacked (range: 100-μm) for comparison. 3D projection of the blood vessel (lectin, red) and neurofilament (green) focusing on the cortex and hippocampus regions, including the cornu ammonis 2 (CA2), dentate gyrus (DG). Scale bars (white, 500-μm; yellow, 50-μm). (c) Quantification of size in nucleus (blue) and blood vessel (red) of the before (unexpanded) and after Paper-MAP in mouse brain. Quantification of total diameter from different ratio combinations. Data are presented as mean ± SD (standard deviation); n = 5 for each experimental group. Illustration of mechanism is visualized by PowerPoint v2016. Application of Paper-MAP to the histopathologic assessment of various rodent and human tissues. We also sought to apply the Paper-MAP method to xenograft and organoid models to demonstrate its feasibility. We generated 100 µm-thick slices of U87MG brain glioma xenografts isolated at either the early (9 days post-transplantation) or middle stages (26 days post-transplantation) of tumor progression (Fig. 7), and performed immunostaining for epidermal growth factor (EGFR), proto-oncogene c-Mer Proto Oncogene Tyrosine Kinase (MERTK), and lectin. We observed a substantial increase in EGFR and c-Mer U87MG xenograft samples processed via Paper-MAP. We also performed Paper-MAP on slices generated from human submandibular gland (SMG) organoids, with subsequent immunostaining for alpha smooth muscle actin (α-SMA), Na-K-Cl cotransporter-1 (NKCC1), cytokeratin 18 (CK18) and aquaporin 5 (AQP5) (Fig. 8). Both tissues were easily cleared and remained intact for successful visualization with confocal microscopy. immunostaining of 5xFAD mouse brain tissues at 3 and 9 months compared to wildtype, processed via Paper-MAP. Enlarged photos (5 × 2 panels, horizontal × vertical tiles) are of hippocampal region. Yellow arrowheads indicate hemolytic plaques. Scale bars (green, 1000-μm; yellow, 100-μm). Quantification of blood vessel length (b), volume (c) and volume/length (d) in 5xFAD and wildtype hippocampus at 3 and 9 months. Quantification of total vessel length from different ratio combinations. Data are presented as mean ± SD (standard deviation); n = 5 for each experimental group. Illustration of mechanism is visualized by PowerPoint v2016. www.nature.com/scientificreports/ Finally, to address the limitations of the original MAP protocol with regards to human biopsies, we performed Paper-MAP on a neurosurgical tumor biopsy, which was successfully cleared and expanded within 1 day. For comparison, we also performed the original MAP protocol, which cleared and expanded tumor tissue over a period of 10 days (Fig. 9a). As shown in Fig. 9b,c, there were no observable differences in the degree of tissue clearance achieved by either protocol, despite the significantly reduced incubation time of Paper-MAP. Immunostaining for epidermal growth factor receptor (EGFR), c-Mer proto oncogene tyrosine kinase, oligodendrocyte transcription factor 2 (Olig2), and lectin (blood vessel) of human brain tumor samples processed via optimized Paper-MAP also demonstrated the successful application of this technique to patient-derived samples.

Discussion
In this study, we introduce Paper-MAP, a modified MAP technique that allows for the rapid clearing and expansion of tissues for three-dimensional ultrastructural imaging, within 2-3 days. During the time of its inception, the original MAP technique significantly advanced the ability to appreciate three-dimensional tissue architecture 9 . Its major limitations, however, included the long duration of time required for tissue processing, as well as the harsh effects of the denaturation and dissociation steps on tissue integrity, which require incubation at high temperatures without pre-treatment with crosslinking reagents to ensure tissue stability.
A major innovation of the Paper-MAP techniques is that they allow for the processing of 100 µm-thick sections as opposed to whole tissue, which enables the conservative use of rare samples that are difficult to obtain, such as patient biopsies. It is highly scalable, allowing for a single sample to be probed for multiple targets during subsequent immunostaining. The use of 100-µm sections as opposed to whole tissues also overcomes the limitation imposed by the 2-mm working distance of most imaging systems, rendering the need for a special, costly super-resolution microscope unnecessary. Other method for adjustment of tissue expending exists. Expansion microscopy (ExM) is a method to expand biological specimens (100-µm to 1-mm thick slices) with protease digestion of a hydrogel-tissue hybrid homogenizes the tissue's mechanical characteristics, and allows four-fold linear expansion 10 . However the protease digestion step in ExM that can causes a loss of proteins, which limits the number of protein structures that can be imaged in the same sample.
Paper-MAP also uses an ammonium persulfate (APS) and tetramethyl ethylenediamine (TEMED)-based solution for embedding as opposed to the V50 cationic azo initiator used in the original MAP protocol, which reduces the time required for tissue embedding from over 3 days to 5 minutes 7,13 . Furthermore, whereas the original MAP protocol uses high denaturation temperatures as well as high concentrations of acrylamide (20-30%) to prevent tissue crosslinking, Paper-MAP involves the direct processing of tissue samples upon fixation in 4% PFA, which improves tissue stability. Fluorescence stability in sample is remains stable at room temperature and 37 °C in all incubation process containing clearing (denaturation) in previous study 13 . Whereas Paper-MAP (or original MAP) requires protein denaturation process of tissue at 95 °C, and it can be that some loss of labeled fluorescence signals. But this problem is possible surmountable problem to additional immunostaining at after Paper-MAP step, and it has no effect on the imaging analysis. To demonstrate the feasibility of Tissue-MAP www.nature.com/scientificreports/ (original MAP), we processed tissues harvested from mouse models of spinal cord injury according to our Tissue-MAP protocol, and performed immunostaining of lectin and DiD-D to visualize vasculature and neuronal networks, respectively. With the Tissue-MAP technique, we were able to successfully clear and expand injured spinal cord tissue such that we could observe blood vessel thinning and reduced neuronal density adjacent to the scar tissue, supporting its use for three-dimensional ultrastructural visualization of intact tissues. Nevertheless, www.nature.com/scientificreports/ Tissue-MAP required long processing times for tissue clearance, expansion, and subsequent immunostaining for super-resolution imaging, over a period of 20 days. We then further optimized our Paper-MAP, which requires even shorter incubation times by removing the incubation step in low acrylamide (A.A) solution for 24 h. Paper-MAP allows for tissue clearance, expansion, and subsequent immunostaining in 2 days, compared to over 20 days for the original Tissue-MAP protocol. Using Paper-MAP, we observed vasculatures and neural structures with immunostaining of lectin dye and neurofilament antibodies in the cortex region, and in cornu ammonis 2 (CA2) and dentate gyrus (DG) of the hippocampus region using confocal microscopy. We also investigated hippocampal vasculature in the 5xFAD mouse model of Alzheimer's disease (AD), compared to that of wildtype littermates. We observed age-dependent progression of disease, from blood vessel thinning and increased microvasculature apparent at 3 months to large, hemolytic plaques at 9 months 14 . Previous studies have suggested a role for blood-derived amyloid-β protein 15 , and while further research is required, our findings point to their potential role in these hemolytic plaques. Nonetheless, these results demonstrate the use of Paper-MAP to visualize disease pathology in a three-dimensional manner. Upon confirming the efficacy of Paper-MAP within the context of Alzheimer's disease, we sought to apply the technique to a variety of pathological conditions, including two mouse models of spinal cord injury and a xenograft model of human glioma in mice. For spinal cord tissues, we specifically investigated angiogenesis, which is known to be defective in the early stages of spinal cord injury (SCI) 16,17 , as well as neuronal damage by immunostaining for lectin and neurofilament, respectively 18 . Consistent with previously established findings, we observed relatively high cell cohesion as well as tapering or disconnected neurofilaments and blood vessels near the scar region of hemisection-induced SCI tissues 19,20 . We also performed immunostaining of neurofilament, GFAP, blood vessels, γ-aminobutyric acid (GABA), parvalbumin, and tyrosine hydroxylase in deletion-induced mouse SCI tissues, which did not demonstrate such cell cohesion in scar tissue. In glioma xenograft-derived tissues, we observed the expression patterns of EGFR and c-Mer, which are known markers for cell proliferation and differentiation in tumors 21,22 .
Finally, we demonstrated the applicability of Paper-MAP to human tissues, including human submandibular gland (SMG) organoids, and more importantly, brain tumor biopsies. Another limitation of the original MAP protocol is that, because a key step involves the whole perfusion of the model organism with Tissue-MAP solution prior to harvesting the organ of interest, it is difficult to apply the technique to human tissues. Paper-MAP does not require whole perfusion and successfully clears tissues fixed in 4% PFA, which makes it amenable to patient derived biopsies. In sections of SMG organoids, we observed the expression patterns of alpha smooth muscle actin (α-SMA), Na-K-Cl cotransporter-1 (NKCC1), cytokeratin 18 (CK18) and aquaporin 5 (AQP5) [23][24][25][26] , and in Figure 7. Application of Paper-MAP for tissue clearing and expansion in mouse brain tumor. Schematic representation of Paper-MAP processing using 100 μm-thick mouse brain tumor sections. Paper-MAP processing of human glioma xenografts upon 9 and 26 days post-transplantation in mice, followed by immunostaining for EGFR (green), c-Mer (green) and lectin (red). Scale bars: green = 2000-μm. Illustration of mechanism is visualized by PowerPoint v2016. www.nature.com/scientificreports/ the sections of human brain tumor biopsies, we assayed for the tumor-specific markers epidermal growth factor receptor (EGFR), c-Mer proto oncogene tyrosine kinase, and oligodendrocyte transcription factor 2 (Olig2) 27,28 .
In conclusion, Paper-MAP represent a significant advancement in the field of tissue clearing, as they allow for the clearing and subsequent visualization of individual tissue sections, in an unprecedented short amount of (c-e) Paper-MAP processing of human SMG organoids and subsequent immunostaining (range: 100-μm) for alpha-SMA (green), CK18 (red or green), NKCC1 (green), AQP5 (red) and DAPI (blue). Images of before (left, fixed in 4% PFA) and after Paper-MAP (right) was taken with the same 10 × and 40 × objective lens and z-stacked (range: 100-μm) for comparison. Scale bars: yellow = 50-μm. Illustration of mechanism is visualized by PowerPoint v2016. www.nature.com/scientificreports/ www.nature.com/scientificreports/ time. Given the demonstrated efficiency, efficacy, and versatility of the technique, Paper-MAP has strong potential for further elucidating the mechanisms underlying complex biological processes, especially with regards to the spatial relationships between cellular and subcellular structures that contribute to these phenomena.

Materials and methods
Human brain tumor samples. All   To induce spinal cord injury, mice were anesthetized with 2% isoflurane, and a dorsal laminectomy was performed at the C5 region of the spinal cord. The dura was removed using microscissors and forceps. In this study, we used two models of spinal cord injury. The lateral hemi-section model involved the use of a microblade in the C5 spinal cord region. In the deletion model, we used a microspatula to remove regions C4-C6 along the dorsal midline blood vessel. In both models, injuries were applied only to the right side of the C5 spinal cord region. Wounds were closed using a 4-0 black silk suture. All mice received 500 μL of sterile saline, cefazolin (25 mg/kg; Bristol Myers Squibb, New York, NY, USA), and buprenorphine (0.05 mg/kg; Reckitt and Colman Pharmaceuticals Inc., VA, USA) for 3 days after surgery. The workflow of for subsequent behavioral tests used to assess the two SCI models is in Supplementary Figs. S1 and S2 online. Mice exhibiting hand cannibalism and/ or sudden death, and those that otherwise could not be evaluated with these behavioral tests, were excluded from the study.
3. Mouse orthotopic xenograft model for brain glioma tumor 6 to 8-week old male athymic nude mice were purchased from Central Lab. Animal Inc. (Seoul, Korea). 2 × 10 5 dissociated U87MG cells (ATCC® HTB-14™; American Type Culture Collection, VA, USA) were implanted into the right frontal lobe of mice at a depth of 4.5 mm using a guide-screw system and Hamilton syringe. Gossypol (40 mg/kg; Sigma-Aldrich Inc., St. Louis, MO, USA) and phenformin (100 mg/kg; Sigma-Aldrich Inc., St. Louis, MO, USA) were orally administered daily. At either 9 or 26 days post-treatment, mice were euthanized and their brain tissues were harvested. If body weight decreased by more than 15% relative to baseline throughout the duration of the experiment, mice were euthanized according to the guidelines of the approved animal protocol. All experiments were performed as previously described 21 .
Behavioral tests of SCI mouse models. Behavioral tests were first performed 2 days post-surgery, and tests were conducted weekly from 1 to 4 weeks after surgery. Forelimb recovery was assessed using the forelimb locomotor rating scale (FLRS) 29 , grip strength measurement (GSM) 30 , forelimb foot fault scoring (FFS) 31 , and the Irvine, Beatties, and Bresnahan (IBB) 32 forelimb recovery scale. Hindlimb recovery was assessed using the Basso mouse locomotor scale (BMS), and hindlimb FFS 29,32,33 . In addition, body weight was measured 1 h before testing.

Forelimb locomotor rating scale (FLRS)
Mice were placed inside a transparent acrylic glass path, and mouse movements were recorded twice from the front and the back of the path by two experienced observers. Forelimb function recovery was assessed on a 17-point scale consisting of joint movement, weight support, stepping, predominant paw position, and toe clearance. The average time mice took to pass through the acrylic glass path was approximately 15 s. Grip strength was measured using a grip strength meter (GSM, TSE Systems; SciPro Inc., London, United Kingdom). Mice were pre-trained three times a week for two weeks. Paper tape was placed on a single forepaw, and the maximum force of grip strength was measured by pulling the tail after the free paw without paper tape caught the GSM bar. Grip strength was tested three times per week for 6 weeks post-injury.

Foot fault scoring (FFS)
The FFS test was performed using a video-recorded ladder rung walking task. The rung walking apparatus was consisted of two Plexiglas walls (70 cm × 15 cm) with 0.12 cm diameter holes at 5 cm intervals. The holes were filled with 8-cm metal bars (diameter 0.1 cm). Foot fault was scored using the following scale: 0 point (total miss), 1 point (deep slip), 2 points (slight slip), 3 points (replacement), 4 points (correction), 5 points (partial placement) and 6 points (correct placement).

Irvine, Beatties, and Bresnahan (IBB) forelimb recovery scale
During the test, mice were given two minutes to eat circle-shaped cereal in a transparent acrylic glass cylinder (10-cm diameter) with glass mirrors on each side. Forelimb recovery was assessed via the following: predominant elbow position, proximal forelimb movement, contact non-volar support, predominant forepaw position, contact volar support, cereal adjustments, wrist movement, contact digit movements, and grasping method. Mice were also scored based on the original IBB scale, which consists of predominant elbow position, forepaw position, cereal adjustments, digit movements, and grasping method. Mice were acclimated to the testing environment 2 weeks prior to surgery, during which they were given a cereal diet within the cylinder used in the test.

Basso mouse locomotor scale (BMS)
Basso mouse locomotor scale (BMS) testing was performed simultaneously with FLRS under similar conditions. After taking two video recordings per animal, FLRS and BMS were verified by video evaluation. In BMS, the scores were evaluated based on 9-point scale consisting of ankle movement, plantar stepping, coordination, paws parallel, trunk stability, and tail movement.

Injured lesion preparation and quantification.
At each experimental endpoint, mice were anesthetized with an overdose of zoletil (Virbac, Carros, France) and rompun (Bayer HealthCare, Leverkusen, Germany). The thorax was exposed, and an incision was made in the right atrium of the heart. Trans-cardiac perfusion was performed with equal volumes of ice-cold 0.1 M PBS and 4% paraformaldehyde (PFA) using a 50 mL syringe. After fixation for 1 day in 4% PFA, the solution was replaced with 30% sucrose in 0. Human salivary gland organoid culture. Human single clonal stem cells were used to establish salivary gland organoids, as previously described 26 . Cells were seeded in a petri dish coated with 1% of Pluronic F127 (Sigma-Aldrich Inc., St. Louis, MO, USA) in Phosphate-buffered saline (PBS) at a density of 40,000 cells/cm 2 . Suspended cells were cultured in low glucose DMEM media (Life Technologies Co., Carlsbad, CA, USA) supplemented with 10% FBS and 100 U/ml streptomycin-penicillin (Invitrogen Inc., Carlsbad, CA, USA). After 7 days of culture, the established salivary gland organoid was post-treated to suit the purpose of the experiment. MAP technique. Tissue clearing, denaturation, and expansion were performed according to previously established protocols 9 , which were specifically optimized for the clearing and expansion of whole intact tissues.

Perfusion of experimental mouse
Mice were anesthetized with 2% isoflurane. Upon opening the thorax, an incision was made in the right atrium of the heart. Trans-cardiac perfusion was performed with equal volumes of ice-cold 0.  Supplementary Table S1 online.
Prior to imaging, tissues were re-expanded in dH 2 O at room temperature for 1-24 h. Intact whole tissue samples were placed on a slide inside a U-shaped Blu-Tack adhesive (Bostik, WI, USA), and covered with a glassbottom Wilco or confocal dish filled with dH 2 O. Samples of Paper-MAP can long term storage in dH 2 O for until a month at room temperature. Fluorescence-labeled samples of Paper-MAP can long term storage for 1-2 months with shrinkage form in refractive index matching solutions (EasyIndex; LifeCanvas Technologies, Cambridge, MA, USA). All clear images were captured using iPhone-X camera (Apple Inc., Cupertino, CA, USA). Neural fiber and blood vessel images of whole spinal cord tissue were obtained with a LaVision Light-sheet Ultramicroscope (LaVision BioTec GmbH, Bielefeld, Germany) at 2.0 × (0.5 NA) magnification. All images of Paper-MAP were captured using a confocal microscope (LSM780 and LSM980; Carl Zeiss, Oberkochen, Germany) at a magnification of 10 × (0.45 NA, 2.0 mm working distance), 20 × (0.8 NA), and 40 × (0.8 NA). Images were processed and analyzed using Zeiss ZEN-2 software (Carl Zeiss, Oberkochen, Germany), and results were processed into three-dimensional images and videos using Imaris v8.0.1 software (Bitplane, Belfast, United Kingdom).
Quantification of vascular architecture. We compared the lengths, density, and volume fraction of the hippocampal vessels between several sub-regions including the three cornu ammonis (CA1, CA2, CA3) regions and dentate gyrus (DG) in wild type and 5xFAD brains. Upon performing immunostaining for lectin, blood vessel length and volume were quantified using Imaris v8.0.1 software. Images were reconstructed with "Surpass view" on the "Surface" icon in the objects toolbar. To analyze the Region of Interest (ROI), 10 ROIs were selected from each sub-region. "Channel" was used as the source channel, and "Absolute Intensity" was used to adjust thresholds. Results were filtered by the number of voxels using the "Classify Surfaces" tab, according to default settings. Quantification was performed using the "Statistics" tool.