Immuno-fluorescent Labeling of Microtubules and Centrosomal Proteins in Ex Vivo Intestinal Tissue and 3D In Vitro Intestinal Organoids

The advent of 3D in vitro organoids that mimic the in vivo tissue architecture and morphogenesis has greatly advanced the ability to study key biological questions in cell and developmental biology. In addition, organoids together with recent technical advances in gene editing and viral gene delivery promises to advance medical research and development of new drugs for treatment of diseases. Organoids grown in vitro in basement matrix provide powerful model systems for studying the behavior and function of various proteins and are well suited for live-imaging of fluorescent-tagged proteins. However, establishing the expression and localization of the endogenous proteins in ex vivo tissue and in in vitro organoids is important to verify the behavior of the tagged proteins. To this end we have developed and modified tissue isolation, fixation, and immuno-labeling protocols for localization of microtubules, centrosomal, and associated proteins in ex vivo intestinal tissue and in in vitro intestinal organoids. The aim was for the fixative to preserve the 3D architecture of the organoids/tissue while also preserving antibody antigenicity and enabling good penetration and clearance of fixative and antibodies. Exposure to cold depolymerizes all but stable microtubules and this was a key factor when modifying the various protocols. We found that increasing the ethylenediaminetetraacetic acid (EDTA) concentration from 3 mM to 30 mM gave efficient detachment of villi and crypts in the small intestine while 3 mM EDTA was sufficient for colonic crypts. The developed formaldehyde/methanol fixation protocol gave very good structural preservation while also preserving antigenicity for effective labeling of microtubules, actin, and the end-binding (EB) proteins. It also worked for the centrosomal protein ninein although the methanol protocol worked more consistently. We further established that fixation and immuno-labeling of microtubules and associated proteins could be achieved with organoids isolated from or remaining within the basement matrix.


Introduction
Formation of epithelia with apico-basal polarity is a fundamental process in development and involves a dramatic reorganization of the microtubules and centrosomal proteins. A radial microtubule array emanating from a centrally located centrosomal microtubule organizing center (MTOC) is prominent in many animal cells and this is well suited for relatively flat cells. In contrast, columnar epithelial cells, such as those of the intestine, assemble non-radial transcellular microtubule arrays that better support the shape and specialized functions of these cells. This dramatic reorganization of the microtubules is achieved by the centrosome moving to the apex and apical non-centrosomal MTOCs (n-MTOCs) forming, which becomes responsible for anchorage of the transcellular microtubules 1,2,3,4,5 .
Much of our knowledge of epithelial differentiation and the associated microtubule reorganization has come from investigations of 2D in vitro cell layers that do not display the in vivo tissue architecture. Development of 3D in vitro organoid cultures, pioneered by Clevers and co-workers 6 , represents a major technological advancement as they mimic in vivo architecture and development. A hierarchy of epithelial differentiation is evident in the intestine; stem cells at the bottom of crypts give rise to immature transit amplifying cells that proliferate and gradually differentiate as they migrate up the crypt onto the small intestinal villus or colonic surface, where they become fully differentiated prior to being shed into the lumen 7 . Importantly, this is replicated in intestinal organoids where cells from the stem cell niche proliferate forming cysts that subsequently generate crypt-like buds with stem cells at the bottom and differentiation gradually progressing towards the cyst region, which becomes villuslike 8 . The intestinal organoid therefore represents a powerful model to study not only microtubule and centrosomal reorganization during epithelial differentiation but numerous other proteins, as well as providing an ideal platform for screening of drugs and food compounds of potential therapeutic benefits 9,10 .
Organoids are well suited for live-imaging of fluorescent-tagged proteins and both knock-in and knock-out organoids can be generated using CRISPR/Cas9 gene editing . NOTE: To view different sections of the small intestine then at this stage separate and treat separately. See Figure 1 and Table 1 for a flow diagram and timings. 2. Flush the contents of the small intestine with PBS using a glass pipette with rubber bulb. 3. Cut open the small intestine with dissecting scissors and place in 15 mL PBS in a Petri dish. 4. Gently wash the intestine in PBS to remove luminal content while not damaging the mucosal surface. Transfer the tissue to a fresh Petri dish and repeat PBS wash. 5. Cut the intestine into 3-5 mm pieces and transfer to a 100 mm Petri dish containing 15 mL of 30 mM EDTA in PBS and incubate for 5 min at RT. Alternatively incubate in 3 mM EDTA in PBS for 30 min. 6. Fraction 1 (Villi isolation): Transfer the intestinal pieces into 10 mL of PBS in a 50 mL conical tube, shake vigorously for 10 s, and pour into a 100 mm Petri dish. Check the fraction for isolated villi under a stereomicroscope. At this stage, mostly villi should be present but this can vary between isolations. NOTE: To prevent isolated villi/crypts sticking to the plastic, Petri dishes and collection tubes should be pre-coated with Fetal Bovine Serum (FBS). Pour FBS into dishes and/or tubes and immediately remove it. See Table 1 for timing guides for each fraction. 7. Transfer the intestinal pieces back into 30 mM EDTA in PBS and incubate for 5 min. During this incubation, collect Fraction 1 in a 15 mL conical tube (precoated with FBS) and centrifuge at 300 x g for 5 min. 8. Fraction 2: Transfer intestinal pieces to a 50 mL conical tube containing 10 mL of PBS, shake vigorously for 20 s, and then pour into a 100 mm Petri dish. Check the fraction under the microscope. Typically, at this stage a mixed culture of villi and crypts are present (Figure 2A). 9. Transfer intestinal pieces back into 30 mM EDTA in PBS and incubate for a further 5 min. During this incubation, pellet Fraction 2 in a 15 mL conical tube (precoated with FBS) at 300 x g for 5 min. Remove supernatant from Fractions 1 and 2 without disturbing the pellet and proceed immediately to fixation (step 3). 12. Fraction 5: Transfer the intestinal pieces into 10 mL of PBS, shake for 20 s, and pour into a 100 mm Petri dish. Check fraction under the microscope. Usually, at this stage, very few crypts are extracted. If more than a few crypts are extracted, then continue with a 6 th fraction. NOTE: If a pure crypt population is desired (for example, for organoid generation), then use a 70 µm cell strainer to remove any intact villi from the crypt-enriched fraction. 13. Collect Fractions 3 -5 in separate 15 mL conical tubes and centrifuge at 300 x g for 5 min. 14. Check the pellets of Fractions 3 -5 and remove the supernatants without disturbing the pellets, and proceed immediately to fixation (Section 3).

Isolation of Intestinal Organoids from Basement Matrix Domes in 24-well Plates
NOTE: The formation of organoids within basement matrix domes has been described elsewhere 12 .
1. Coat wells with basement matrix dome according to the manufacturer's instructions.
2. Wash wells containing basement matrix domes with 500 µL of PBS. 3. Add cold 250 µL of cell recovery solution (4 °C) to each well (see Table of Materials). NOTE: The cold cell recovery solution will depolymerize all but stable microtubules. 4. Scrape the basement matrix domes using a P1000 micropipette and carefully pipette up and down throughout the well to breakup and remove the basement matrix from plastic. 5. Collect the supernatant in 1.5 mL low binding microcentrifuge tubes. 6. Invert the 1.5 mL low binding tube several times and check under the microscope whether the organoids have been isolated and are free moving and not in clumps. Hold the tube under a stereomicroscope and view under low (50X) magnification. 7. Pellet the organoids by centrifugation at 1,000 x g for 5 min at RT. 8. Remove the recovery reagent and proceed immediately to fixation (step 3).

Fixation of Isolated Intestinal Tissue and Organoids
4. Use the micropipette to transfer the mixture of crypts/villi/organoids in the mounting media solution, and dispense in a line along the center of a microscope slide. NOTE: The line should not be longer than the cover-glass that is to be used. Check using a stereomicroscope whether the crypts/villi/ organoids are well spread on slide and not clumped. 5. Carefully place a cover-glass over the top; avoid generating bubbles. 6. Place the glass slide in a slide book and store in a fridge overnight for the mounting media to set before analyzing on a confocal microscope.
NOTE: For mouse antibody staining in mouse tissue, use the commercial immunofluorescence kit (see Table of Materials). Replace the blocking step (Section 4) with a 10 min block with protein blocking solution followed by a 1 h incubation with the blocking reagent that comes with the kit. Then at step 5.8 (in the middle of washing) add a 1 h incubation with fluorescence signal enhancer reagent. Then use the kit's reagent in fluorescent dilutant (other secondary antibodies can also be used in combination with this reagent). Incubate as above and proceed from step 6 with the rest of the protocol.

Fixation and Immuno-labeling of Organoids Within Basement Matrix
NOTE: Organoids destined for fixation and immuno-labeling while remaining within the basement matrix were generated in basement matrix domes on top of round glass coverslips in a 24-well plate (one dome per well). The organoid basement matrix domes were processed within the 24-well plate by the addition and removal of the various solutions. 13. Using forceps, remove the glass coverslip with the basement matrix dome and place on a slide with the basement matrix dome facing upwards; add a few drops of mounting media and then put a glass coverslip on top. 14. Place the glass-slide in a slide book and leave in fridge overnight for the mounting media to set.

Isolation of intestinal tissue for immuno-labeling
The described tissue isolation protocols for colon and small intestine were optimized for preservation and immuno-labeling of microtubules and associated proteins, but not for stem cell viability and organoid generation (Figure 1 and Table 1). The aim was to generate crypt and villus factions that were as clean (devoid of mucus and other tissue) as possible, while minimizing exposure to EDTA and cold to preserve structure and prevent depolymerization of microtubules with ice cold solutions, which induce depolymerization of all but stable microtubules. Figure 2 shows examples of images of Fractions 2 and 3 from isolated small intestinal tissue, with Fraction 2 containing a mixture of both villi and crypts (Figure 2A, B), while Fraction 3 contains mainly crypts (Figure 2C, D).

Fixation and immuno-labeling of isolated intestinal tissue
The individual or combined fractions were then processed for fixation and immuno-labeled through a series of steps including fixation, detergent, blocking, antibody, and washing solutions, before re-suspending the final villi/crypts pellet in mounting media, transferring to slides, and covering with glass coverslips. The crypts and villi were then imaged on a confocal microscope.
Good preservation and labeling of microtubules and actin in both villi and crypts was achieved by the following: a combination of formaldehyde/ methanol fixation at -20 °C, repeated washing in PBS containing 0.1% detergent and 1% serum and blocking in PBS with 0.1% detergent and 10% serum, followed by overnight incubation at 4 °C in primary antibodies and then 2 h in secondary antibodies at room temperature ( Figure  3). Formaldehyde/methanol fixation also worked well for labeling +TIPs such as the EBs and CLIP-170 in isolated crypts and villi (Figure 4). EB3 accumulations at the plus-end of microtubules (known as comets) were evident in crypts (Figure 4A), while association along the lattice of stable microtubules could be seen in villi samples (Figure 4C). Distinct localization of CLIP-170 and p150 Glued (subunit of dynactin) was clearly evident at the apical n-MTOCs in isolated villi ( Figure 4B). Fixation with the formaldehyde/methanol protocol did not consistently work for ninein localization in isolated intestinal tissue using our Pep3 antibody against mouse ninein. However, methanol fixation at -20 °C followed by the same washing and blocking solutions as for formaldehyde/methanol gave very good localization of ninein within isolated crypts and villi ( Figure 5; reference 8 ). Interestingly, while ninein is concentrated at the apical centrosomes some accumulation at the cell base was evident in some cells within isolated crypts (Figure 5). Whether this is due to non-specific labeling or a consequence of the isolation procedure delaying fixation (and thus affecting preservation) will need further investigation. However, methanol-fixed (-20 °C) cryostat sections of villi (see Figure 3bi in 8 ) also revealed ninein at the cell base in some cells suggesting that ninein may also associate with a basal population of microtubules.

Isolation of intestinal tissue
Isolation of small intestinal crypts and villi and colonic crypts involves exposing the mucosal surface, treatment with EDTA solution to loosen cell contacts, fractionation (shaking), and centrifugation. The presented intestinal villi/crypt isolation protocol has been modified from Belshaw et al. and Whitehead et al. 17,18 Exposing the mucosal surface We have experimented with a number of approaches to expose the mucosal surface of the intestinal tract in the development of this procedure. A classic approach is to evert (turn inside out) the tube, usually in segments about 100 mm long, using a metal rod that is caught in a fold of the tissue at one end and then the remaining tube slid over the tube 19 . For mouse tissue, a metal rod (2.4 mm diameter) with rounded ends is ideal. This approach has the benefit of expanding the mucosal surface allowing better access to PBS and EDTA. We initially used this approach but moved to cutting the tube into short lengths (about 5 cm) and opening each section with dissecting scissors as this proved easier. This approach is appropriate if only a few intestines are required; but if more animals were to be used in an experiment then a purpose-built device for cutting open the tube longitudinally, as described by Yoneda et al. 14 would be more efficient. Initially we used 3 mM EDTA in PBS and relatively long incubation times of up to 60 min to loosen the mucosal surface from the underlying tissue 17,18 . At this concentration of EDTA we found an incubation time of 30 min was sufficient to loosen crypts from mouse colon. However, for the small intestine crypt/villus isolation we tried using more concentrated EDTA for a shorter time, which proved to be an efficient approach. All subsequent work was undertaken with tissue extracted using the 30 mM EDTA technique generating relevant fractions for villi or crypts. For crypts, we normally pooled fractions 3 -5 before fixing but it is important to check whether these are the appropriate fractions as the timings will depend on a number of factors such as position along the intestinal tract, age of mouse, inflammation, previous diet, etc. Similarly, the length of time the tissue needs to be shaken following the EDTA treatment to be effective may vary under different conditions. The result is isolated tissue fractions containing a mixture of villi and crypts or mainly either villi or crypts (Figure 2). As there are no villi in the colon, the crypt extraction may be achievable in one step by shaking the tissue in the tube for 30 s. These fractions can then be fixed and processed for immuno-labeling.

Isolation of intestinal organoids from basement matrix
Isolation of organoids from basement matrix domes can be achieved by using cell recovery solution. The solution works by depolymerizing the gelled basement matrix but the temperature needs to be 2 -8 °C. A note of caution is that dynamic microtubules may not be preserved. Thus, for immuno-labeling of dynamic microtubules and +TIPs such as the EBs cell, recovery from basement matrix prior to fixation is not recommended. However, most of the microtubules in the differentiating organoid cells are relatively stable and these were preserved (Figure 6). It also worked well for immuno-labeling of centrosomal and junctional proteins as well as cell markers.

Fixation protocols
Formaldehyde (freshly made from PFA) is a relatively rapid-acting fixative that forms reversible cross-links and 4% PFA works well for example, in immuno-labeling microtubules and gamma-tubulin and staining actin filaments with Phalloidin. More dilute PFA solutions such as 1% worked well for immuno-labeling for example, with the stem cell markers Lgr5 and Paneth cell marker CD24 within the crypt stem cell niche, while higher concentrations of PFA did not work.
The addition of glutaraldehyde gives better preservation of microtubules and the so called PHEMO fixation which consists of a mixture of 3.7% PFA, 0.05% glutaraldehyde and 0.5% detergent in PHEMO buffer (68 mM PIPES, 25 mM HEPES, 15 mM EGTA and 3 mM MgCl2) 2 gives excellent preservation of microtubules without compromising antigenicity. It also works well for immuno-labeling gamma-tubulin, β-catenin, and E-cadherin, and staining actin filaments with phalloidin. However, in 3D tissue and organoid cultures, the PHEMO fixation produced inconsistent results and was therefore not used.
Methanol is a coagulant fixative that gives relatively good penetration and tends to preserve antigenicity. Fixation with 100% methanol (-20 °C) introduces some shrinkage, gives moderate morphology preservation, and works for microtubules, +TIPs, and many centrosomal antibodies including ninein in 2D cell cultures. However, some organoids collapsed when using this fixation method. In addition, penetration of antibodies through entire crypts, villi, or organoids was initially a problem but the addition of 0.1% detergent to the wash solution and prolonged washing achieved better results.
A combination of formaldehyde and methanol had previously been used by Rogers et al. 20 to immuno-label EB1 in Drosophila. A fixation protocol based on a mixture of formaldehyde and methanol was therefore developed for intestinal tissue and organoids based on 3% formaldehyde and 97% methanol chilled to -20 °C, but omitting the 5 mM sodium carbonate from the mixture that was used by Rogers et al. 20 In addition, samples were fixed in the freezer at -20 °C. This worked particularly well for immuno-labeling +TIPs, such as CLIP-170 and the EBs, but also proved excellent for fixing and immuno-labeling microtubules and actin within tissue and 3D organoids. Very good structural preservation was evident and antigenicity was preserved for several cytoskeletal and associated proteins as well as centrosomal proteins such as gamma-tubulin and ninein, although labeling for ninein worked more consistently with methanol fixation.

Disclosures
The authors declare no competing financial interests.