Detection and Quantitation of Label-Retaining Cells in Mouse Incisors using a 3D Reconstruction Approach after Tissue Clearing

Summary

The mouse incisor contains valuable label-retaining cells in its stem cell niche. We have a novel way to unbiasedly detect and quantify the label-retaining cells; our study used EdU labeling and a 3D reconstruction approach after PEGASOS tissue clearing of the mandible.

Abstract

The murine incisor is an organ that grows continuously throughout the lifespan of the mouse. The epithelial and mesenchymal stem cells residing in the proximal tissues of incisors give rise to progeny that will differentiate into ameloblasts, odontoblasts, and pulp fibroblasts. These cells are crucial in supporting the sustained turnover of incisor tissues, making the murine incisor an excellent model for studying the homeostasis of adult stem cells. Stem cells are believed to contain long-living quiescent cells that can be labeled by nucleotide analogs such as 5-ethynyl-2´-deoxyuridine (EdU). The cells retain this label over time and are accordingly named label-retaining cells (LRCs). Approaches for visualizing LRCs in vivo provide a robust tool for monitoring stem cell homeostasis. In this study, we described a method for visualizing and analyzing LRCs. Our innovative approach features LRCs in mouse incisors after tissue clearing and whole-mount EdU staining followed by confocal microscopy and a 3-dimensional (3D) reconstruction with the imaging software. This method enables 3D imaging acquisition and non-biased quantitation compared to traditional LRCs analysis on sectioned slides.

Introduction

The continuously growing mouse incisor is an excellent model for studying adult stem cells¹. The epithelial (labial and lingual cervical loop) and mesenchymal stem cells (between the labial and lingual cervical loop) that reside on the proximal side of the incisor differentiate into ameloblasts, odontoblasts, and dental pulp cells. This unique process provides a source of cells for compensation of tissue loss and turnover². Though several stem cell markers such as Sox2, Gli1, Thy1/CD90, Bmi1, etc. have been identified in vivo for the subsets of adult stem cells in mouse incisors, they are inadequate in representing the stem cell populations when used alone^1,3,4,5. Visualizing long-living quiescent cells by nucleotide analog DNA labeling and retention could provide unbiased detection for most subsets of adult stem cells⁶. Further, this approach is useful among many stem cell identification methods⁷ for understanding cell behavior and the homeostasis of dental stem cell populations^3,8. While the dividing stem cells would lose their DNA labeling after a considerable chase, the putative quiescent stem cells retain their DNA label, deeming them label-retaining cells (LRCs)⁶. DNA labeling and retention by non-dividing stem cells will mark and locate the putative adult stem cells in their niches.

Over the past years, thymidine analog 5-bromo-2′-deoxyuridine (BrdU) labeling replaced the cumbersome, time-consuming, and high-resolution microscopy incompatible 3H-thymidine DNA labeling method for cell proliferation assays^6,9. In recent years, the 5-ethynyl-2´-deoxyuridine (EdU) labeling technique has been increasingly used over BrdU. This pattern emerged due to several reasons. First, the BrdU method is slow and labor-intensive. User conditions are variable, and it is unable to preserve the ultrastructure in specimens (due to DNA denaturation). Likewise, the BrdU method loses the antigenicity of cells, thus making it inefficient for downstream functional analyses and assays such as the co-localization experiments and in vivo stem cell transplantation^{3,7,9,10,11,12}. BrdU is also a teratogen, which is not suitable for labeling LRCs in embryonic development⁶. Also, the BrdU method is inefficient when used in the whole-mount specimens. The disadvantages are low penetration of antibodies in the deep part of specimens or the requirement of a long antibody incubation period for deep penetration¹³. EdU labeling escapes the steps of denaturing specimens, thus preserving the ultrastructure. This feature is advantageous for downstream functional analyses such as co-localization experiments and stem cell transplantation^11,12. Also, EdU labeling is highly sensitive and rapid; specimen penetration is high due to the use of rapidly absorbed and smaller-sized fluorescent azides for detecting EdU labels through a Cu(I)-catalyzed [3 + 2] cycloaddition reaction ("click" chemistry)¹⁴.

Another increasingly applied DNA labeling method is the use of engineered transgenic mice. These mice express histone 2B green fluorescent protein (H2B-GFP) controlled by a tetracycline-responsive regulator element^5,14. After feeding mice with tetracycline chow/water for a 4-week to 4-month chasing period, the GFP fluorescence will diminish in cycling cells and only LRCs retain the fluorescence⁶. The advantage of this method is that the labeled LRCs can be isolated and remain viable for cell culture or downstream functional analyses^6,7. Some studies reported inaccurate labeling of quiescent stem cells when chased for long-term use. This result was due to a leaky background expression from the H2B-GFP strain and not the appropriate tetracycline-regulated response¹⁵.

Moreover, most literature in the past used the LRCs detection mainly on sectioned slides, which are two dimensional and often erroneously biased in showing the accurate location and number of LRCs. The approach displayed incorrect angles for sections of complex tissue structures¹⁶. The other method was to obtain 3D images from serial sections and perform post-image reconstructions. These steps were inaccurate because of image distortion from variations in each serial section due to compressed or stretched sections, resulting in missing information^16,17,18. The method was also laborious and time-consuming.

To facilitate the whole-mount imaging of LRCs, samples need to be made clear while the fluorescence be well maintained. Current tissue clearing techniques can be classified into three major categories: organic solvent-based tissue clearing techniques, aqueous reagent-based tissue clearing techniques, and hydrogel-based tissue clearing techniques^17,19. The polyethylene glycol (PEG)-associated solvent system (PEGASOS) has been recently developed. This approach renders nearly all types of tissues transparent and preserves endogenous fluorescence, including hard tissues such as bone and teeth²⁰. The PEGASOS method has advantages over other tissue clearing methods, especially in clearing tooth and bone materials. Most other methods could only partially clear hard tissues, have long processing times, or require costly reagents²¹. Also, the PEGASOS method can efficiently preserve endogenous fluorescence over other methods.

This literature led us to create a new method for cell study. We combined the LRCs detection advantages of EdU labeling with the most superior 3D whole-mount imaging of tissue-cleared specimens; samples were processed with advanced polyethylene glycol (PEG)-associated solvent system (PEGASOS) tissue clearing technique¹⁵. Hard tissue transparency enabled us to reconstruct the 3D signal of LRCs fluorescence in vivo without breaking the teeth or mandible, creating a more accurate way to visualize and quantify LRCs.

In this study, we provide an innovative guide to visualize LRCs in the mouse incisor. We made a 3D visual approach to determine the location and quantity of LRCs within the mouse incisor stem cell niche. This project used EdU labeling, PEGASOS tissue clearing techniques, and confocal microscopy. Our method of EdU labeling the LRCs on whole-mount tissue and the use of a cleared and transparent specimen overcomes both the limitations of traditional sectioned slides and other disadvantageous DNA labeling methods. Thus, our technique will be suitable for studies on stem cell homeostasis requiring LRCs detection, especially on hard tissues. The protocol can be equally advantageous to those focusing on stem cell homeostasis in other tissues and organs.

Protocol

All methods described here have been approved by the Institutional Animal Care and Use Committee (IACUC) for Texas A&M University College of Dentistry.

1. Preparation of the EdU labeling cocktail

1. Prepare the cocktail stock solutions as described in Table 1.
2. Prepare the EdU labeling cocktail as described in Table 1.

2. Preparation of the mice and EdU solution

1. Prepare the EdU solution. Dissolve EdU in DMSO at 20 mg/mL and store in a -20 °C freezer.
   CAUTION: EdU is a mutagen. Users should wear appropriate personal protective equipment (PPE) when handling this substance.
2. Inject 5-day-old C57BL/6J mice with the prepared EdU solution intraperitoneally at 3 µL/g body weight. Mice must receive the solution once a day for 7 consecutive days. Next, they undergo a chase period for 6 weeks before being harvested for analysis (Figure 1).
   ​CAUTION: Be careful to avoid needle pricking when injecting EdU into the mice. After injecting EdU for 7 consecutive days, place the mice in a new cage. After the mice are in their new home, dispose of the bedding in the old cage using appropriate biohazard rules.

3. Sample preparation by trans-cardiac perfusion (postnatal 53-day-old mice after the 6-week chase period)

NOTE: The mouse liver should turn pale after successful trans-cardiac perfusion.

1. Anesthetize mice with an intraperitoneal injection. They need to be given a combination of xylazine and ketamine anesthetics (Xylazine 10-12.5 mg/kg; Ketamine, 80-100 mg/kg body weight).
2. Wait for ~10 min until the mouse no longer responds to painful stimuli (such as paw pinch reflexes).
3. Place the mice in a supine position stabilized on Styrofoam support with needles in each paw.
4. Expose the chest cavity by opening up the overlying skin using tweezers and dissecting scissors.
5. Cut open the diaphragm using sharp scissors. Take care not to pierce the heart.
6. Cut through the ribcage, grabbing the base of the sternum with clamp scissors to expose the heart.
7. Transfer the mouse to the perfusion stage near the circulation pump in a fume hood.
8. Insert the 25 G needle (from the tubing filled with phosphate-buffered saline (PBS)/4% paraformaldehyde (PFA) solution) into the apex of the left ventricle. Take care to keep the tip of the needle in the lumen of the ventricle.
   CAUTION: PFA is moderately toxic and a probable carcinogen. Use PPE to handle PFA and prepare the solution under a chemical fume hood. Waste PFA requires proper disposal per institution guidelines.
9. Cut the right ventricle using a sharp scissor immediately after inserting the needle into the left ventricle. This step allows for blood to flow out of the mouse and drain into the collecting dish.
10. Perfuse the samples with 50 mL of PBS solution at a flow rate of 7 mL/min.
11. After PBS perfusion, switch the stopcock to allow for a flow of 20 mL of 4% PFA solution at the same flow rate of 5 mL/min.
12. While the circulation pump is still on, remove the needle from the left ventricle.
    ​NOTE: The mouse is now fixed for tissue collection. To preserve the fluorescence, light exposure should be avoided whenever possible throughout the entire protocol. While the tissue is processing, the 50 mL conical tube with the tissue samples may be covered with aluminum foil.

4. Tissue clearing of the mandible using the PEGASOS technique

NOTE: A 15 mL or 50 mL conical tube as per the volume of tissues can be used to keep the samples ready for treatment in each step. Samples are processed at 37 °C shakers (~100 rpm) from steps 4.2 to 4.7. Use polypropylene-based plastic containers that are resistant to organic solvents to avoid melting the plastic. Alternatively, glassware can be used.

CAUTION: The PEGASOS tissue clearing technique uses toxic solutions such as Quadrol, polyethylene glycol (PEG), benzyl benzoate (BB), MMA500, etc. Appropriate PPE is required to avoid potential exposures.

1. Dissect mandibles and further fix samples in 4% PFA. Keep them at room temperature overnight.
2. Remove the muscles from the mandibles and immerse them in 0.5 M EDTA solution (pH 8.0) to decalcify for 4 days. During this period, perform a daily medium change on a 37 °C shaker.
   NOTE: It is desirable to remove the muscles from the sample as much as possible. This precaution will simplify the imaging process and reduce the autofluorescence from the muscles.
3. Decolorize the mandibles with a 25% Quadrol solution (v/v in H₂O) on a 37 °C shaker for one day to remove the remaining blood heme.
4. Complete whole mount EdU staining.
   1. Prepare a fresh EdU labeling cocktail according to Table 1.
      NOTE: The EdU labeling cocktail needs to be made fresh each time with freshly prepared ascorbate solution added for obtaining optimum EdU labeling results.
   2. Rinse the mandibles with PBST for 30 min on a 37 °C shaker.
   3. Rinse the mandibles with PBS three times for 3 min each time.
   4. Incubate the mandibles in a freshly made EdU labeling cocktail on a 37 °C shaker overnight.
   5. Rinse the mandibles with PBS three times for 3 min each time.
5. Perform serial de-lipidation in each of the following solutions for 6 h on 37 °C:
   30% tert-butanol (tB) solution: 75% v/v H₂O, 22% v/v tB, and 3% v/v Quadrol.
   50% tB solution: 50% v/v H₂O, 47% v/v tB, and 3% v/v Quadrol.
   70% tB solution: 30% v/v H₂O, 67% v/v tB, and 3% v/v Quadrol
   NOTE: The de-lipidation step is critical. De-lipidation can be done between 4-6 h in 30% and 50% tB solution and for 1 day in 70% tB solution.
6. Dehydrate the mandibles in tB-PEG solution for 2 days with the following daily medium change: 75% tB, 22% v/v polyethylene glycol methyl-ether methacrylate average MMA500, and 3% v/v Quadrol at 37 °C.
7. Clear the mandible in benzyl benzoate (BB)-PEG clearing medium to obtain refractive index matching: 75% v/v BB, 22% v/v PEG-MMA500, and 3% Quadrol until transparency is achieved.
   NOTE: We have information on refractive index matching. The inorganic and organic components in tissues such as water, lipids, proteins, minerals, organelles, etc. have a different refractive index (RI) in the range of 1.33 to 1.55. This RI mismatch between the tissue components hinders the imaging process. Transparency can be achieved by eliminating RI mismatch within the tissue. Immersing the tissue-cleared samples in the clearing medium of BB-PEG is advised. This step will give the uniform internal RI of ~1.54 (called RI matching) to the whole specimen, easing the imaging process.
8. Preserve the mandibles in the BB-PEG tissue clearing medium at room temperature for storage and imaging.
   ​NOTE: It is good to complete the imaging as soon as possible to avoid the gradual loss of fluorescent reporters. However, the PEGASOS method will keep fluorescence reporters well preserved even after several weeks of storage^20,21. For long-term storage, the samples can also be stored at 4°C.

5. Confocal imaging of the tissue-cleared mandible

1. Mount the tissue cleared whole mandible on brand cavity slides in the BB-PEG clearing medium and cover with glass cover slides. Avoid any bubbles.
2. Perform image acquisition with a suitable laser scanning confocal microscope. Select Acquire and set the image acquisition parameters at 512 x 512 pixel resolution and 400 Hz acquisition speed. These settings are found in the acquisition mode menu of the software that controls the microscope.
3. In the panel for fluorescent excitation, activate the laser required (TRITC) to optimally excite fluorophores (the probe we used for LRCs visualization has fluorescent properties such as Cy3 with excitation at 548 nm and emission at 563 nm of laser wavelength).
4. In the panel for fluorescent detection, move the slider to select the wavelengths that will be measured (for example, between 555 to 625 nm for a Cy3-like fluorophore).
5. Place the mounted mandible on the microscope stage to visualize and find the sample using white light and a lower magnification objective (2-10x).
6. Add a drop of immersion oil with a refractive index of 1.52 on top of the coverslip to match the refractive index of the glass coverslip and objective.
   NOTE: Match the refractive index of objectives, imaging media, and tissue as closely as possible to minimize the introduction of optical distortions during image acquisition.
7. Turn on the fluorescent lamps and choose an appropriate magnification objective to image regions of interest. We used a 20x objective lens with a numerical aperture of 0.9 with a working distance of 1.95 mm. A longer working distance lens may be required when imaging larger tissue samples in the confocal microscope. Alternatively, imaging can be done easily using a light-sheet microscope for larger tissue samples.
8. On the confocal imaging software, press the Live button to start live-scan mode. To maximize the dynamic range of the images and avoid pixel saturation, adjust the gain and laser intensity for the channel selected.
9. Identify the field of view to visualize the entire region of the mandible in the X and Y dimensions. Set the upper and lower Z-stage acquisition parameters that accurately cover the volume of the stem cell niche in the mice incisor. Use a Z-step size of 2 µm or as per your requirements. The newer versions of confocal microscopes should be able to automatically acquire z-stacks of the various overlapping regions.
10. Acquire and save Z-stack image files in .lif format.
    ​NOTE: The .lif flies can be converted into .tiff files and used for other 3D analysis software.

6. Image processing, 3D image reconstruction, and quantification of label-retaining cells by creating a surface, segmentation of the Region of Interest (ROI), masking the ROI, and creating spots

NOTE: We used Imaris (Bitplane 9.0.1) for image processing and 3D reconstruction, but similar image processing steps can be conducted using other software suites (e.g., ImageJ/Fiji, 3D slicer, Avizo, Arivis, Amira, etc.).

1. In the imaging software, click the Arena button, then choose Image. Select the original .lif file and choose merged file, the images will be imported into the imaging software.
   NOTE: Alternatively, use an imaging software file converter to convert Z stack image file to the native format file type for the software. For example: convert the .lif files into .ims files using the Imaris File Converter and open the .ims files in the imaging software. Converting data to the native file format of the software will allow rapid file conversion and minimize errors.
2. Double click the imported file to open the image data set.
3. Visit the Display Adjustment panel,click the name of the channel. It is usually labeled as Red in default mode.
   1. In the Image Properties window, click on the Mapped Color tab. In the Color Table File option, choose Fire Flow, and then click on OK. The display color is normally chosen to distinguish the background fluorescence and the positive fluorescence from the sample.
4. On the upper left side of the screen in the Scene – Properties pane, click on the blue icon labeled Add New Surface. Unselect the Volume icon which will remove most of the background fluorescence and the positive fluorescence are distinctly visible (see Step 6.4.2).
   1. Start the process of manual segmentation of region of interest (ROI) by clicking on the Create tab and selecting the Skip Automatic Creation, Edit Manually tab. In the manual surface creation wizard, click on the Contour tab to choose the orientation (XY, or YZ, or XZ) based on the samples to easily contour the ROI. Here, we choose XY orientation.
   2. Next, make sure the pointer menu in the right corner is in Select mode. Adjust the Slice Position to locate the ROI in the tissue. You will notice that most of the background fluorescence have been removed with distinct positive fluorescence as mentioned in Step 6.4. Click on the Draw button and draw a tentative region of interest. Select the Visibility option to none to avoid interference from other areas than ROI.
   3. Repeat step 6.4.2. slice by slice to draw the entire region of interest.
   4. After the ROI drawing is finished, click on Create Surface to get a 3D geometry of the ROI.
   5. Click on the Edit button in the manual surface creation wizard and select Mask All.
   6. In the popped-out Mask Channel window, select Channel 1 in the channel selection options. Afterward, select Duplicate Channel Before Applying the Mask. In the mask settings, select Constant Inside/Outside and set the voxel outside surface to 0.
   7. In the Scene-Properties pane, deselect the Surface Object and select the Volume Object. In the Display Adjustment, deselect the original "Red” channel and select the Masked Red channel and adjust the color intensities to get the desired 3D-rendered and positively labeled fluorescence ROI.
5. Create a new spot object. Begin by clicking on the Add on Spots icon to quantify the 3D-rendered LRCs by creating 3D spots that comparably overlap with 3D-rendered LRCs. Use the bottom arrows to move between steps in the spot creation wizard.
   1. In the spot creation wizard, there will be four sequential steps of instructions to create spots automatically with the software. Make sure the Segment Only a Region of Interest option is selected in the first step. Click on the Next icon.
   2. In the second step, there will be an "XYZ 3D box". Adjust the XYZ box to cover the ROI. Click on the Next icon.
   3. In the third step, go to the Source Channel options. Choose the masked red channel; input the estimated XY spot diameter comparable to fit and overlap the sample fluorescence dots. Click on the Next icon.
   4. In the fourth step, add the "Quality" filter type and adjust the lower intensity threshold by moving the sliding window across the 'Quality' histogram until the majority of identifiable LRCs are labeled.
      NOTE: The segmentation and therefore statistical readouts (example: number of cells) very much depend on the intensity threshold values. Be careful to precisely set the appropriate threshold values.
   5. Click on the Finish button and select the Statistic button. Select the Overall tab to find the value for the total number of spots in the image.
   6. Export statistics by clicking on the Tab Display to File button and receive the results in an .xls file.
   7. Use a suitable diagram to present the quantification results.

Results

After EdU labeling and the PEGASOS tissue clearing process (Figure 2), the transparent mandible was obtained as shown in our image (Figure 3B). We compared the modified sample to a normal mandible without a tissue clearing process (Figure 3A). The transparent mandible (Figure 3B) with EdU labeling was subjected to confocal imaging. We focused on the incisor apex showing the stem cell niche as shown in (...

Discussion

Multiple doses of injections (BrdU, EdU) are usually used on growing neonatal mice to label proliferating cells as much as possible^1,6,13. The chasing period is considered a critical step regarding the renewal rate of tissues^6,13. The mouse incisor renews itself around every month. This trait allows researchers to set the chasing period to 4 weeks or longer

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
0.5 M EDTA	Sigma Aldrcih	E9884
20 × Objective/NA 0.9	Leica	507702
50 mL Falcon Centrifuge Tubes	Falcon	352070
BD PrecisionGlide Needle	BD	REF 305111
Bezyl benzoate (BB)	Sigma Aldrcih	409529
Bitplane 9.0.1	Imaris
BRAND cavity slides	Millipore Sigma	BR475505
C57BL/6J mice	Jackson Laboratory	Strain #:000664
Circulation Pump	VWR	23609-170
CuSO4	Sigma Aldrcih	451657
DMSO	Sigma Aldrcih	D8418
EdU	Carboynth	NE08701
Heparin	Miiilipore Sigma	H3149
Imaging System	Olympus	DP27
LAS X Software	Leica
Olympus Stereo Microscope	Olympus	SZX16
Paraformaldehye	Sigma Aldrich	P6148
PBS	Sigma Aldrich	P4417
PEGMMA500	Sigma Aldrich	447943
Quadrol	Sigma Aldrich	122262
Sodium Ascorbate	Sigma Aldrich	11140
Sulfa-Cyanine 3 Azide	Lumiprobe	D1330
TBS-10X	Cell Signaling Technology	12498
TCS SP8 Confocal Microscope	Leica
tert-butanol (tB)	Sigma Aldrich	360538
Triton X-100	Sigma Aldrich	X100