In Vivo Imaging Reveals Existence of Crypt Fission and Fusion in Adult Mouse Intestine

Abstract

The intestinal epithelium is a repetitive sheet of crypt and villus units with stem cells at the bottom of the crypts. During postnatal development, crypts multiply via fission, generating 2 daughter crypts from 1 parental crypt. In the adult intestine, crypt fission is observed at a low frequency. Using intravital microscopy in Lgr5^{EGFP-Ires-CreERT2} mice, we monitored individual crypt dynamics over multiple days with single-cell resolution. We discovered the existence of crypt fusion, an almost exact reverse phenomenon of crypt fission, in which 2 crypts fuse into 1 daughter crypt. Examining 819 crypts in 4 mice, we found that 3.5% ± 0.6% of all crypts were in the process of fission, whereas 4.1 ± 0.9% of all crypts were undergoing crypt fusion. As counteracting processes, crypt fission and fusion could regulate crypt numbers during the lifetime of a mouse. Identifying the mechanisms that regulate rates of crypt fission and fusion could provide insights into intestinal adaptation to altered environmental conditions and disease pathogenesis.

Editor's Notes.

Background and Context

Stem cells located in intestinal crypts constantly drive self-renewal of the intestinal epithelium. Duplications of these crypt structures (fission) facilitate among others regenerative responses.

New Findings

Using multiday in vivo imaging in mice, the authors identified a new phenomenon in the intestine of adult mice: crypt fusion, an almost exact reverse phenomenon of crypt fission.

Limitations

Future research needs to determine the frequencies and kinetics of crypt fission and fusion to better understand homeostasis, aging and pathogenesis of the intestine in mouse and men.

Impact

Stem cells accumulate mutations that drive aging and cancer. Crypt fission and fusion influences the number of intestinal stem cells, and therefore impact aging and cancer.

Crypts of Lieberkühn are central elements of the self-renewing nature of the intestine because they harbor active cycling stem cells at their base and progenitor cells along the flanks. During postnatal development, crypts are dynamic structures that undergo multiple rounds of replication via a process called crypt fission to accommodate the elongation of the intestinal tract.¹ In general, this process occurs through bifurcation of individual crypts where branching starts at the base and elongates toward the villus in a zipper-like fashion. This bifurcation is thought to be completed within 1 week.² In addition to elongation of the intestinal tract, this process has been shown to play an essential role in the regenerative response after irradiation and tissue resection.^{3, 4}

Interestingly, generation of new crypts does not completely disappear in adulthood, where fission remains continuously present.^{2, 5} Intriguingly however, the relative high number of crypt fission events^{2, 5} seems to be at odds with the observation that small intestinal length only slightly increases during aging, while both small intestinal width and crypt density remain constant (Supplementary Figure 1). In addition, analyses of crypt ancestry by Cre-induced lineage tracing⁶ or mutation-induced marker systems⁷ also indicate that fewer crypt fission events have occurred than the frequency of crypt branching events suggest. Previously, the concept of the crypt cycle has been proposed to describe the continuous growth and bifurcation of crypts analogous to the cell cycle.⁸ Mechanisms that counterbalance the continuous production of new crypts have been postulated, including crypt death, but direct evidence has never been reported.

Supplementary Figure 1.

Dimensions of small intestine during aging. (A) Frequency of all crypts with a bifurcation phenotype in mice of different ages, counted in whole mounts of distal intestine. (B) Length and (C) width of small intestine from mice of different ages. (D) Crypt density, measured as average distance between centers of neighboring crypts, in mice of different ages.

Recently, we developed intravital microscopy (IVM) techniques that enable us to monitor intestinal stem cells, crypt size, composition, and dynamics over multiple days with single-cell resolution.⁹ To identify mechanisms that have the ability to counterbalance the continuous production of new crypts, the small intestine of mice was positioned behind an abdominal imaging window and imaged over multiple consecutive days. Using Lgr5^{EGFP-Ires-CreERT2} knock-in mice (See Supplementary Materials and Methods in supplementary material section), we took advantage of the fluorescent marking of intestinal stem cells at the crypt bottom, ie, Lgr5⁺ crypt base columnar cells. Following crypt morphology over time using IVM, we indeed observed crypts undergoing fission in line with previous observations.^{5, 10}

Surprisingly, in addition to previously described crypt fission, we observed crypt fusion events where 2 independent crypts merge together to form 1 daughter crypt with 1 central lumen (Figure 1A). Intriguingly, based on the evolving crypt morphology during the process, a fusion event seems to be an almost exact reversal phenomenon of a crypt fission process (Figure 1B,C). As such, crypt fusion has the potential to act as a counterbalancing mechanism for the continuous birth of crypts via crypt fission.

Figure 1.

Crypt dynamics includes both crypt fission and crypt fusion. (A) Crypt fusion event where 2 separate crypts (day 0) merged into 1 crypt (day 4). Overview panels (middle) show crypt pattern to confirm crypt identity. Outer panels show enlargement of imaging planes at center and border of stem cell zone. (B) Representative examples of crypt fusion and (C) fission where a single event is followed over 5 consecutive days by IVM. Arrows point at separate lumen. Dotted lines indicate outlines of Lgr5^EGFP+ (green) crypts and lumen. Cartoons illustrate process at beginning (left) and end (right). Scale bars: 20 μm.

To confirm that crypt fusion involves 2 autonomous and independent crypts rather than being a reversal of incomplete crypt fission, we crossed Lgr5^{EGFP-ires-CreERT2} mice with the Cre-reporter mouse strain LSL-tdTomato. In this mouse model, the LSL-tdTomato allele recombines at very low background levels (0.86% ± 0.31% of all Lgr5⁺ crypts, 4 mice; Supplementary Figure 2A), thereby occasionally giving rise to individual, fully tdTomato-labeled crypts as a result of neutral drift.^{11, 12} As expected, we found tdTomato-labeled crypts with an 8-shaped crypt circumference that is representative of the intermediate state of either crypt fission or fusion (Figure 2A). The distribution pattern of tdTomato-labelled cells within 8-shaped crypts allowed us to discriminate between these 2 processes (Supplementary Figure 2B). Crypt fission events were scored when 8-shaped crypts were fully labeled because the chance of independent labeling events in 2 neighboring crypts before fusion could be neglected because of the low labeling frequency. In contrast, 8-shaped crypts in which labeling was restricted to one half were scored as crypt fusion events because fissions of non-clonal crypts were not observed (for further explanation see Supplementary Figure 2B). After examining 819 fully labeled crypts (4 mice), we found that 3.5% ± 0.6% of all labeled crypts were in the process of crypt fission, while 4.1% ± 0.9% of all labeled crypts were in the process of crypt fusion (Figure 2A,B; Supplementary Figure 2C).

Supplementary Figure 2.

Discrimination between crypt fission and fusion. (A) Representative image of whole mount sample of a field of Lgr5^EGFP+ crypts (green) and a single tdTomato⁺ crypt (red). Scale bar: 100 μm. (B) Schematic diagram illustrating the possible processes (fission and fusion) that can yield the 2 distribution patterns of tdTomato⁺ cells over 2 branches of an ‘intermediate’ 8-shaped crypt. Number of observations of α and β pattern are indicated. Crypt fission contributes equally to both labeling patterns. Because α pattern is not observed, we deduced that crypt fusion is responsible for the 33 scored events of the β pattern. (C) Images of all 33 observed fusion events. Scale bar: 20 μm.

Figure 2.

Fusion events of 2 independent crypts. (A) IVM images of 5 consecutive days reveals crypt fusion between tdTomato⁺ (red) and non-labelled crypt. Cartoons illustrate process at beginning (left) and end (right). Dotted lines indicate outlines of Lgr5^EGFP+ (green) crypts and lumen. Asterisks indicate competing cells within neutral drift. Scale bar: 20 μm. (B) Quantification of fission and fusion events in fixed whole mount samples of proximal and distal intestine (819 tdTomato⁺ crypts, 4 mice). Symbols indicate individual mice. Representative examples of crypt fission (left) and fusion (right) are depicted below. (C) Schematic representation of crypt cycle that includes crypt fission and fusion.

Next, using IVM we monitored a substantial amount of sporadically tdTomato-marked crypts and identified a fusion event between 2 independent Lgr5^EGFP+ crypts (Figure 2A). Interestingly, 24 hours after central lumen formation, tdTomato⁺ cells started to intermingle with non-marked cells, indicating that both stem cell pools are now united and participate within the same neutral drift toward clonality.^{11, 12}

Our ability to monitor evolving crypt morphology over time provided the opportunity to identify the existence of crypt fusion during homeostatic conditions in adulthood. Because crypt fission and fusion occur at approximately similar frequency, it can be speculated that, as counteracting mechanisms, they are involved in regulating crypt numbers. However, many variables concerning fission and fusion are poorly understood, underscoring why both frequencies are not exactly in line with the slight increase of intestinal length during aging. For instance, minor deviations in duration of both processes or differences in frequencies per region along the gastrointestinal tract or during aging might explain the discrepancy. Moreover, other non-discovered counterbalancing mechanisms cannot be excluded.

Although opposite processes, crypt fission and fusion are morphologically nearly identical, explaining why crypt fusion has not been identified with conventional analyses of fixed tissue. Identification of immunohistochemistry markers to discriminate between fission and fusion would facilitate studies on human samples. One way to identify such markers would be by combining laser capture microdissection with genomic approaches. Indeed, laser capture microdissection on bifurcating human crypts has already been performed to study methylation patterns.^{13, 14} Intriguingly, adjacent crypts in human colon, and even 2 arms of a bifurcating crypt, can be as dissimilar to one another as 2 unrelated distantly located crypts.^{14, 15} The presence of crypt fusions could very well explain these counterintuitive observations.

Crypt fusion exists in adulthood during homeostasis and is a counterbalancing process for the continuous birth of new crypts (Figure 2C). It is well known that crypt fission underlies the regenerative response of the intestinal epithelium. Identifying the mechanisms that modulate the rates of crypt fission and fusion is of high interest to understand intestinal adaptation to altered environmental conditions in health and disease.

Supplementary Materials

Materials and Methods

Mice

Lgr5^{EGFP-ires-CreERT2} and Lgr5^{EGFP-ires-CreERT2}:LSL-tdTomato male and female mice (mixed background between C57BL/6 and 129/Ola) were housed under standard laboratory conditions and received standard laboratory chow and water ad libitum. Random (double) heterozygous mice were used for the experiments. Male and female mice between 12–14 weeks old were used for IVM. For whole mount imaging intestines from 8–79-week-old male and female mice were used. Background recombination frequency was determined in 4 mice. Quantification of crypt fission and crypt fusion in whole mount samples was based on 819 fully labeled crypts in 4 mice. All experiments were carried out in accordance with the guidelines of the Animal Welfare Committee of the Royal Netherlands Academy of Arts and Sciences.

Whole Mount Preparation

To prepare whole mounts, the small intestine was harvested and directly put on ice. It was flushed 3 times with ice-cold PBSO and cut open along its length. The villi were removed using a cover glass and the tissue was washed in ice-cold PBSO. After fixing the tissue for 30 min in 4% formaldehyde solution (w/v) (Klinipath), it was mounted between 2 coverslips using Vectashield HardSet Antifade Mounting Medium (Vector Laboratories). Crypts were imaged from the bottom using the same equipment and settings as for intravital microscopy described below.

Abdominal Imaging Window Surgery

The abdominal imaging window (AIW) surgery was performed as previously described.⁹ In short, before surgery buprenorphine (100 ug/kg mouse; Temgesic, BD Pharmaceutical Systems) was administered via a subcutaneous injection. All surgical procedures were performed under anesthesia via 2% isoflurane (v/v) inhalation. At the start of the surgery, the skin of the mouse was shaved and disinfected with 70% (v/v) ethanol and a left lateral flank incision was made through skin and peritoneum of the mouse. Next, a purse string suture was placed along the wound edge. The ileum was taken out of the abdominal cavity and placed on top of a disinfected AIW (>1 h in 70% [v/v] ethanol) that was positioned glass-side down. The mesenterium was fixed to the cover glass using Cyanoacrylate Glue (Pattex) and CyGel (BioStatus Limited) was added on top of the ileum to reduce peristaltic movement. After allowing the Ultra Gel and CyGel to dry for 1 min, the AIW, together with the ilium, was inverted and placed into the mouse, ensuring that the skin and abdominal wall were placed into the groove of the AIW. Next, the suture was tightened to secure the AIW into the mouse. After surgery, the mice were closely monitored daily for reactivity, behavior, appearance, and defecation. The mice were provided with food and water ad libitum.

Equipment and Settings

Intravital microscopy was performed as previously described.⁹ An inverted Leica TCS SP5 AOBS 2-photon microscope was used with a chameleon Ti:Sapphire pumped optical parametric oscillator (Coherent) equipped with 4 non-descanned detectors (NDDs) and a ×25 (HCX IRAPO NA0.95 WD 2.5 mm) water objective. The NDDs collect the wavelengths: NDD1 (HyD1): 555–680 nm, NDD2 (HyD2): 505–550 nm, NDD3 (PMT1): 455–505 nm, NDD4 (PMT2): <455 nm. Scanning was performed at 940 nm wavelength. Re-identification of the same crypts over multiple days was accomplished by storing xy coordinates of different positions using the ‘multiple position’ function in the LAS-AF software and using the vasculature and the typical patchy Lgr5⁺ crypt pattern as visual landmarks.

Multi-day Crypt Imaging

Following the AIW surgery, mice were kept under anesthesia with isoflurane (1% [v/v]) for imaging and placed in a custom-designed imaging box as previously described.⁹ After each imaging session mice were allowed to recover before being placed back in their cage. The following 3–4 days, mice were imaged daily for a maximum of 4 h. The climate chamber around the microscope was kept at 36°C and body temperature of the animals was monitored using a rectal probe. The patchy pattern of the Lgr5 knock-in allele, in combination with specific landmarks such as blood vessels, enabled us to discriminate individual crypts that could be repeatedly identified over consecutive days. After imaging, the acquired z-stacks were further processed and analyzed using basic functions in ImageJ software.