Colonic stem cell data are consistent with the immortal model of stem cell division under non-random strand segregation

Objectives:  Colonic stem cells are thought to reside towards the base of crypts of the colon, but their numbers and proliferation mechanisms are not well characterized. A defining property of stem cells is that they are able to divide asymmetrically, but it is not known whether they always divide asymmetrically (immortal model) or whether there are occasional symmetrical divisions (stochastic model). By measuring diversity of methylation patterns in colon crypt samples, a recent study found evidence in favour of the stochastic model, assuming random segregation of stem cell DNA strands during cell division. Here, the effect of preferential segregation of the template strand is considered to be consistent with the 'immortal strand hypothesis', and explore the effect on conclusions of previously published results.
Materials and methods:  For a sample of crypts, it is shown how, under the immortal model, to calculate mean and variance of the number of unique methylation patterns allowing for non-random strand segregation and compare them with those observed.
Results:  The calculated mean and variance are consistent with an immortal model that incorporates non-random strand segregation for a range of stem cell numbers and levels of preferential strand segregation.
Conclusions:  Allowing for preferential strand segregation considerably alters previously published conclusions relating to stem cell numbers and turnover mechanisms. Evidence in favour of the stochastic model may not be as strong as previously thought.     

Role of heterozygous APC mutation in niche succession and initiation of colorectal cancer--a computational study

Mutations in the adenomatous polyposis coli (APC) gene are found in most colorectal cancers. They cause constitutive activation of proliferative pathways when both alleles of the gene are mutated. However studies on individuals with familial adenomatous polyposis (FAP) have shown that a single mutated APC allele can also create changes in the precancerous colon crypt, like increased number of stem cells, increased crypt fission, greater variability of DNA methylation patterns, and higher somatic mutation rates. In this paper, using a computational model of colon crypt dynamics, we evolve and investigate a hypothesis on the effect of heterozygous APC mutation that explains these different observations. Based on previous reports and the results from the computational model we propose the hypothesis that heterozygous APC mutation has the effect of increasing the chances for a stem cell to divide symmetrically, producing two stem cell daughters. We incorporate this hypothesis into the model and perform simulation experiments to investigate the consequences of the hypothesis. Simulations show that this hypothesis links together the changes in FAP crypts observed in previous studies. The simulations also show that an APC(+/-) stem cell gets selective advantages for dominating the crypt and progressing to cancer. This explains why most colon cancers are initiated by APC mutation. The results could have implications for preventing or retarding the onset of colon cancer in people with inherited or acquired mutation of one APC allele. Experimental validation of the hypothesis as well as investigation into the molecular mechanisms of this effect may therefore be worth undertaking.     

An enteroendocrine cell-based model for a quiescent intestinal stem cell niche

We have shown that the kinetics of conversion of intestinal crypt cell populations to a partially or wholly mutant phenotype are consistent with a model in which each crypt contains an infrequently dividing 'deep' stem cell that is the progenitor of several more frequently dividing 'proximate' stem cells. An assumption of our model is that each deep stem cell exists in a growth inhibitory niche. We have used information from the literature to develop a model for a quiescent intestinal stem cell niche. This niche is postulated to be primarily defined by an enteroendocrine cell type that maintains stem cell quiescence by secretion of growth inhibitory peptides such as somatostatin and guanylin/uroguanylin. Consistent with this model, there is evidence that the proteins postulated as defining a growth-inhibitory stem cell niche can act as intestinal tumour suppressors. Confirmation that a growth-inhibitory niche does exist would have important implications for our understanding of intestinal homeostasis and tumorigenesis.     

Overlapping DNA Methylation Dynamics in Mouse Intestinal Cell Differentiation and Early Stages of Malignant Progression

Mouse models of intestinal crypt cell differentiation and tumorigenesis have been used to characterize the molecular mechanisms underlying both processes. DNA methylation is a key epigenetic mark and plays an important role in cell identity and differentiation programs and cancer. To get insights into the dynamics of cell differentiation and malignant transformation we have compared the DNA methylation profiles along the mouse small intestine crypt and early stages of tumorigenesis. Genome-scale analysis of DNA methylation together with microarray gene expression have been applied to compare intestinal crypt stem cells (EphB2^high), differentiated cells (EphB2^negative), Apc^Min/+ adenomas and the corresponding non-tumor adjacent tissue, together with small and large intestine samples and the colon cancer cell line CT26. Compared with late stages, small intestine crypt differentiation and early stages of tumorigenesis display few and relatively small changes in DNA methylation. Hypermethylated loci are largely shared by the two processes and affect the proximities of promoter and enhancer regions, with enrichment in genes associated with the intestinal stem cell signature and the PRC2 complex. The hypermethylation is progressive, with minute levels in differentiated cells, as compared with intestinal stem cells, and reaching full methylation in advanced stages. Hypomethylation shows different signatures in differentiation and cancer and is already present in the non-tumor tissue adjacent to the adenomas in Apc^Min/+ mice, but at lower levels than advanced cancers. This study provides a reference framework to decipher the mechanisms driving mouse intestinal tumorigenesis and also the human counterpart.     

Mouse embryonic stem cells induce targeted DNA demethylation within human MAGE-A1 transgenes

Human tumor development is often associated with a DNA demethylation process. This results in the activation of germline-specific genes, such as MAGE-A1, which rely on DNA methylation for repression in somatic tissues. Here, we searched to identify a cell line possessing ongoing DNA demethylation activity targeted to MAGE-A1. We first assessed MAGE-A1-expressing human tumor cell lines, by evaluating their ability to induce demethylation of MAGE-A1 transgenes that were methylated in vitro before transfection. All cell lines lacked such activity, suggesting that MAGE-A1 hypomethylation in tumors results from a past demethylation event. We then turned to mouse embryonic stem (mES) cells, which are characterized by a high level of methylation plasticity. Interestingly, in vitro methylated MAGE-A1 transgenes became demethylated after transfection into mES cells. Demethylation was targeted to the 5’-region of MAGE-A1, and was strongly reduced at mutated MAGE-A1 transgenes exhibiting impaired promoter activity. Our results indicate that mES cells induce demethylation of MAGE-A1, and represent therefore a valuable system to study this tumor-related process.     

Intestinal crypt properties fit a model that incorporates replicative ageing and deep and proximate stem cells

A model of intestinal crypt organization is suggested based on the assumption that stem cells have a finite replicative life span. The model assumes the existence in a crypt of a quiescent ('deep') stem cell and a few more actively cycling ('proximate') stem cells. Monte Carlo computer simulation of published intestinal crypt mutagenesis data is used to test the model. The results of the simulation indicate that stabilization of the crypt mutant phenotype following treatment with external mutagen is consistent with a stem cell replicative life span of about 40 divisions for mouse colon and 90-100 divisions for mouse small intestine, corresponding to a deep stem cell cycle time of about 3.9 and 8.5 weeks for colon and small intestine, respectively. Simulation of the data obtained for human colorectal crypts suggests that the proximate stem cell cycle time is about 80 h, assuming a replicative life span of 50-150 divisions, and that the deep stem cell divides approximately every 30 weeks.     

A Notch positive feedback in the intestinal stem cell niche is essential for stem cell self‐renewal

The intestinal epithelium is the fastest regenerative tissue in the body, fueled by fast‐cycling stem cells. The number and identity of these dividing and migrating stem cells are maintained by a mosaic pattern at the base of the crypt. How the underlying regulatory scheme manages this dynamic stem cell niche is not entirely clear. We stimulated intestinal organoids with Notch ligands and inhibitors and discovered that intestinal stem cells employ a positive feedback mechanism via direct Notch binding to the second intron of the Notch1 gene. Inactivation of the positive feedback by CRISPR/Cas9 mutation of the binding sequence alters the mosaic stem cell niche pattern and hinders regeneration in organoids. Dynamical system analysis and agent‐based multiscale stochastic modeling suggest that the positive feedback enhances the robustness of Notch‐mediated niche patterning. This study highlights the importance of feedback mechanisms in spatiotemporal control of the stem cell niche.     

Biased competition between Lgr5 intestinal stem cells driven by oncogenic mutation induces clonal expansion

The concept of ‘field cancerization’ describes the clonal expansion of genetically altered, but morphologically normal cells that predisposes a tissue to cancer development. Here, we demonstrate that biased stem cell competition in the mouse small intestine can initiate the expansion of such clones. We quantitatively analyze how the activation of oncogenic K-ras in individual Lgr5⁺ stem cells accelerates their cell division rate and creates a biased drift towards crypt clonality. K-ras mutant crypts then clonally expand within the epithelium through enhanced crypt fission, which distributes the existing Paneth cell niche over the two new crypts. Thus, an unequal competition between wild-type and mutant intestinal stem cells initiates a biased drift that leads to the clonal expansion of crypts carrying oncogenic mutations.     

A stochastic branching model with formation of subunits applied to the growth of intestinal crypts

The intestinal epithelium is one of the most rapidly regenerating tissues in mammals. Cell production takes place in the intestinal crypts which contain about 250 cells. Only a minority of 1-60 proliferating cells are able to maintain a crypt over a long period of time. However, so far attempts to identify these stem cells were unsuccessful. Therefore, little is known about their cellular growth and selfmaintenance properties. On the other hand, the crypts appear to exhibit a life cycle which starts by fission of existing crypts and ends by fission or extinction. Data on these processes have recently become available. Here, we demonstrate how these data on the life cycle of the macroscopic crypt structure can be used to derive a quantitative model of the microscopic process of stem cell growth. The model assumptions are: (1) stem cells undergo a time independent supracritical Markovian branching process (Galton-Watson process); (2) a crypt divides if the number of stem cells exceeds a given threshold and the stem cells are distributed to both daughter crypts according to binomial statistics; (3) the size of the crypt is proportional to the stem cell number. This model combining two different stochastic branching processes describes a new class of processes whose stationary stability and asymptotic behavior are examined. This model should be applicable to various growth processes with formation of subunits (e.g. population growth with formation of colonies in biology, ecology and sociology). Comparison with crypt data shows that intestinal stem cells have a probability of over 0.8 of dividing asymmetrically and that the threshold number should be 8 or larger.     

DIURNAL VARIATION IN PROLIFERATIVE COMPARTMENTS AND THEIR RELATION TO CRYPTOGENIC CELLS IN THE MOUSE COLON

The depth of the crypts in mouse descending colon varied diurnally, between twenty-six cells at 24.00 hours and thirty-eight cells at 12.00 hours. Cell loss from the colon was greatest immediately before the maximum faeces production, at the beginning of the dark period. The labelling index of the colon also changed, from 9% at 20.00 hours to 16% at 12.00 hours. The greatest variation in labelling index occurred at the top of the zone of proliferative cells, between the ninth and eighteenth cell position up the crypt. In this region a synchronized cohort of about forty cells apparently entered S phase once a day. Although the length of the proliferative zone doubled at 12.00 hours, that of the non-proliferative zone remained fairly constant all day. The number of cryptogenic cells per crypt was estimated by comparing single and split-dose X-ray survival curves. This gave a mean value of two cryptogenic cells per crypt. Crypts rarely regenerated from the base after irradiation. The cryptogenic cells probably lay between cell positions Nos 9 and 18 up the crypt and probably did not function as stem cells in the normal crypt.     

Activation-induced deaminase is critical for the establishment of DNA methylation patterns prior to the germinal center reaction

Activation-induced deaminase (AID) initiates antibody diversification in germinal center B cells by deaminating cytosines, leading to somatic hypermutation and class-switch recombination. Loss-of-function mutations in AID lead to hyper-IgM syndrome type 2 (HIGM2), a rare human primary antibody deficiency. AID-mediated deamination has been proposed as leading to active demethylation of 5-methycytosines in the DNA, although evidence both supports and casts doubt on such a role. In this study, using whole-genome bisulfite sequencing of HIGM2 B cells, we investigated direct AID involvement in active DNA demethylation. HIGM2 naïve and memory B cells both display widespread DNA methylation alterations, of which ∼25% are attributable to active DNA demethylation. For genes that undergo active demethylation that is impaired in HIGM2 individuals, our analysis indicates that AID is not directly involved. We demonstrate that the widespread alterations in the DNA methylation and expression profiles of HIGM2 naïve B cells result from premature overstimulation of the B-cell receptor prior to the germinal center reaction. Our data support a role for AID in B cell central tolerance in preventing the expansion of autoreactive cell clones, affecting the correct establishment of DNA methylation patterns.     

DNA methylation and demethylation link the properties of mesenchymal stem cells: Regeneration and immunomodulation

Mesenchymal stem cells (MSCs) are a heterogeneous population that can be isolated from various tissues, including bone marrow, adipose tissue, umbilical cord blood, and craniofacial tissue. MSCs have attracted increasingly more attention over the years due to their regenerative capacity and function in immunomodulation. The foundation of tissue regeneration is the potential of cells to differentiate into multiple cell lineages and give rise to multiple tissue types. In addition,the immunoregulatory function of MSCs has provided insights into therapeutic treatments for immune-mediated diseases. DNA methylation and demethylation are important epigenetic mechanisms that have been shown to modulate embryonic stem cell maintenance, proliferation, differentiation and apoptosis by activating or suppressing a number of genes. In most studies, DNA hypermethylation is associated with gene suppression, while hypomethylation or demethylation is associated with gene activation. The dynamic balance of DNA methylation and demethylation is required for normal mammalian development and inhibits the onset of abnormal phenotypes. However, the exact role of DNA methylation and demethylation in MSC-based tissue regeneration and immunomodulation requires further investigation. In this review, we discuss how DNA methylation and demethylation function in multi-lineage cell differentiation and immunomodulation of MSCs based on previously published work. Furthermore, we discuss the implications of the role of DNA methylation and demethylation in MSCs for the treatment of metabolic or immune-related diseases.     

Heterogeneity in imprinted methylation patterns of pluripotent embryonic germ cells derived from pre-migratory mouse germ cells

Pluripotent stem cells, termed embryonic germ (EG) cells, have been generated from both human and mouse primordial germ cells (PGCs). Like embryonic stem (ES) cells, EG cells have the potential to differentiate into all germ layer derivatives and may also be important for any future clinical applications. The development of PGCs in vivo is accompanied by major epigenetic changes including DNA demethylation and imprint erasure. We have investigated the DNA methylation pattern of several imprinted genes and repetitive elements in mouse EG cell lines before and after differentiation. Analysed cell lines were derived soon after PGC specification, “early”, in comparison with EG cells derived after PGC colonisation of the genital ridge, “late” and embryonic stem (ES) cell lines, derived from the inner cell mass (ICM). Early EG cell lines showed strikingly heterogeneous DNA methylation patterns, in contrast to the uniformity of methylation pattern seen in somatic cells (control), late EG cell and ES cell lines. We also observed that all analysed XX cell lines exhibited less methylation than XY. We suggest that this heterogeneity may reflect the changes in DNA methylation taking place in the germ cell lineage soon after specification.     

Epithelial cell kinetics in the descending colon of the rat

Epithelial cell kinetics were investigated in the descending colon of the rat. The number of cells per crypt was found to be approximately 625, with 33 cells per cell column and 19 cell columns per crypt circumference. The growth fraction of the colonic crypt was 0.42, and proliferating cells were situated largely in the lower half of the crypt. The cell cycle time was 50.5 h, with values for the G1, S and G2 phases of 40.0, 7.6 and 2.9 h respectively. Cell migration studies showed that it took 60-72 h for a cell to migrate from the upper border of the proliferative cell compartment in the crypt to the luminal surface of the colon. Data were also obtained from continuous labelling with tritiated thymidine and from studying the circadian rhythm of proliferative activity, which suggest that the cells in the bottom of the crypt may constitute a separate, more slowly cycling (stem)cell compartment.     

Antiproliferative effects of DNA methyltransferase 3B depletion are not associated with DNA demethylation

Silencing of genes by hypermethylation contributes to cancer progression and has been shown to occur with increased frequency at specific genomic loci. However, the precise mechanisms underlying the establishment and maintenance of aberrant methylation marks are still elusive. The de novo DNA methyltransferase 3B (DNMT3B) has been suggested to play an important role in the generation of cancer-specific methylation patterns. Previous studies have shown that a reduction of DNMT3B protein levels induces antiproliferative effects in cancer cells that were attributed to the demethylation and reactivation of tumor suppressor genes. However, methylation changes have not been analyzed in detail yet. Using RNA interference we reduced DNMT3B protein levels in colon cancer cell lines. Our results confirm that depletion of DNMT3B specifically reduced the proliferation rate of DNMT3B-overexpressing colon cancer cell lines. However, genome-scale DNA methylation profiling failed to reveal methylation changes at putative DNMT3B target genes, even in the complete absence of DNMT3B. These results show that DNMT3B is dispensable for the maintenance of aberrant DNA methylation patterns in human colon cancer cells and they have important implications for the development of targeted DNA methyltransferase inhibitors as epigenetic cancer drugs.