Crypt Fission Peaks Early During Infancy and Crypt Hyperplasia Broadly Peaks During Infancy and Childhood in the Small Intestine of Humans

Abstract

Background: 

Postnatal growth of the small intestine occurs by crypt hyperplasia and by the less recognised mechanism of crypt fission. How the small intestine grows is largely extrapolated from animals and is poorly described in humans.

Aim: 

To investigate crypt fission and crypt hyperplasia as mechanisms of intestinal growth in humans.

Patients and Methods: 

Proximal intestinal samples were taken from 3 neonates at surgical anastomosis, and duodenal biopsies were taken at endoscopy from 16 infants (mean age 0.7, range 0.3–1.7 years), 14 children (mean age 7.9, range 2.4–16.2 years), and 39 adults. Morphometric measures of villous area, crypt length (measure of crypt hyperplasia), and percentage of bifid crypts (measure of crypt fission) were assessed by a microdissection technique.

Results: 

Mean crypt fission rates in neonates, infants, children, and adults were 7.8%, 15%, 4.9%, and 1.7%, respectively. In particular, crypt fission peaked at 18% in 5 infants from 6 to 12 months of age. Mean crypt length was 123 μm in neonates, 287 μm in infants, 277 μm in children, and 209 μm in adults. Thus, crypt hyperplasia had a broad peak during infancy and childhood.

Conclusions: 

We conclude that crypt fission was present predominantly during infancy, and crypt hyperplasia occurred during both infancy and childhood.

Growth and development of the small intestine has been extensively studied for more than 50 years, but mainly in animals with few reported studies in humans ^(1–3). How the small intestine grows in humans remains uncertain and is largely extrapolated from animal studies. It is important to understand and define the development of the small intestine in humans because the potential application is to treat critically ill neonates, infants, children, and adults with reduced intestinal length and surface area impairing intestinal function. We previously addressed this problem in a study of 74 infants and 77 adults that showed that crypt hyperplasia was associated with intestinal growth during infancy and childhood ⁽⁴⁾. However, a subgroup of 5 exclusively breast-fed infants who were less than 6 months old did not demonstrate crypt hyperplasia, suggesting that other additional mechanisms may be important ⁽⁵⁾. Nearly all of the studies have investigated potential growth factors that promote crypt hyperplasia ^(2,3). It is recognised that there remains an elusive intestinal growth factor that could potentially be used to treat impaired intestinal surface area in premature neonates or to treat short bowel syndrome in children and adults ⁽⁶⁾.

Another mechanism of intestinal growth is by longitudinal division of crypts, where the crypt unzips from the base and is initiated by division of stem cells located at 4 and 5 cell positions from the crypt base ^(7,8). This process is termed crypt fission (also called branching). Crypt fission was described 35 years ago by Clarke ⁽⁹⁾ in the rat, and the description was extended by St Clair and Osborne ⁽¹⁰⁾, who showed that crypt fission was associated with an increase in the number of intestinal crypts. Crypt fission has been described in a single human neonate being treated with epidermal growth factor ⁽¹¹⁾. Cheng and Bjerknes ⁽¹²⁾ observed that crypt branching as a percentage of total crypts was 30% in infant mice and 5% to 10% in adult mice. In rats, we found that crypt fission peaked early in infant animals at 10.5% and decreased after weaning to adult levels of 1.5% ⁽¹³⁾. Nevertheless, crypt fission is not widely acknowledged as a means of intestinal growth. It also has recently been alluded to in the specialised literature of experimental developmental biology in reference to pathological aberrant crypt branching in murine models of juvenile intestinal polyposis ⁽¹⁴⁾, and in humans and mice with mutations in the adenomatous polyposis coli gene ⁽¹⁵⁾. It was inferred from these studies that crypt fission occurs physiologically, but this process has not been further investigated in humans. What has not been shown is at what age crypt fission is present in humans. This may be relevant in targeting treatment to enhance intestinal growth.

The purpose of this study was to investigate the time course of crypt fission and of crypt hyperplasia with age. We recruited subjects who were undergoing endoscopy for reflux esophagitis to study normal growth of the small intestine because reflux esophagitis is not associated with disease of the small intestine.

PATIENTS AND METHODS

Neonates were recruited with parental consent from those undergoing surgical resection and reanastomosis. Infants, children, and adults were recruited from those having endoscopy for investigation of possible reflux esophagitis that would not affect the small intestine. Informed consent was obtained from parents or guardians for children and from adolescents older than 14 years. Duodenal biopsies were obtained for research from the second part of the duodenum and placed in Clarke fixative (3:1 ethanol-acetic acid). Biopsy data from 6 infants previously reported were included in this study ⁽⁵⁾. The adult data comes from a common database being used for several unrelated studies.

Microdissection of Biopsies and Morphometry

A microdissection technique was used, in which a small fragment of intestine is stained with Feulgen reagent before being returned to water ^(16,17). The serosa is removed and rows of villi and attached crypts are cut from the mucosal fragment using a cataract knife. These are mounted in 45% acetic acid as a wet film on a microscope slide and measurements made of villous length, apical and basal villous width, crypt length, and percentage of bifid crypts. Villous area was calculated as a trapezoid approximation ⁽¹⁶⁾. Villous area and crypt length are proportional to the number of villous and crypt epithelial cells, respectively ⁽¹⁸⁾. This is because villi change in length and width, but crypts change in length only. The percentage of bifid crypts is the index of crypt fission and crypt length is the measure of crypt hyperplasia.

Ethics

This study was approved by the Human Ethics Committees of The Queen Elizabeth Hospital in Adelaide, Women's and Children's Hospital in Adelaide, Royal Children's Hospital in Melbourne, and Sydney Children's Hospital, and was carried out according to the National Statement on Ethical Conduct in Research Involving Humans (1999) of the National Health and Medical Research Council of Australia, and according to the Declaration of Helsinki.

Statistics

Group means were compared using Peritz multiple comparison F test ⁽¹⁹⁾.

RESULTS

Patients

We recruited 3 neonates (<4 weeks of age), 16 infants (<2 years of age), 14 children, and 39 adults. One neonate had surgery for intestinal perforation from necrotizing enterocolitis, another had a resection of a stricture following necrotizing enterocolitis, and a third had surgery for gastroschisis. The mean ages (range) of neonates, infants, children, and adults were respectively 2.4 (0.9–4) weeks, 0.7 (0.3–1.7) years, 7.9 (2.4–16.2) years, and 46 (20–80) years. Infants had mixed breast- or bottle feeding or bottle feeding alone. There were no exclusively breast-fed infants. There were no untoward events.

Changes in Villous Area and Crypt Length

Intestinal morphometry was available from 3 neonates, 6 infants from 1 to 6 months of age, 5 infants from 6 to 12 months of age, 5 infants from 12 to 24 months of age, 14 children, and 39 adults (Fig. 1). Mean villous area peaked in infants of 1 to 6 months and was significantly higher compared with that of neonates (P = 0.039) or of adults (P = 0.012) (Fig. 1). Mean crypt length was 123 μm in neonates, 287 μm in infants, 277 μm in children, and 209 μm in adults (Fig. 1). Thus, crypt length had a broad peak during infancy and childhood.

FIG. 1: Changes in villous area and crypt length with age in neonates, infants, children, and adults. Data are given as mean ± standard error.

Changes in Crypt Bifid Index With Age

Mean crypt fission rates in neonates, infants, children, and adults were 7.8%, 15%, 4.9%, and 1.7%, respectively (Fig. 2). In particular, crypt fission peaked at 18% in 5 infants from 6 to 12 months of age. Mean bifid percentage was 4.6-fold higher in neonates than adults (P < 0.00001), and peaked at 11.2-fold higher in infants at 6 to 12 months compared with adults (P < 0.0001) (Fig. 2). Crypt fission declined in infants at 12 to 24 months but was still elevated at 2.9-fold higher in children than in adults (P < 0.0001). Figure 3 shows what are presumed to be consecutive stages of crypt fission from the same biopsy of 1 individual. There is initial indenting in the crypt base followed by longitudinal division of the crypt (Fig. 3).

FIG. 2: Changes in intestinal crypt fission are seen with age in neonates, infants, children, and adults. Crypt fission was measured as mean percentage of bifid crypts. Data are given as mean ± standard error.

FIG. 3: Illustration of stages of crypt fission in microdissected intestinal crypts. The early stage has a small indent at the crypt base (left), followed by unzipping of the crypt longitudinally (center), and duplication of the crypt (right). Original magnification ×400.

DISCUSSION

Our study has shown that both crypt fission and crypt hyperplasia contributed to postnatal growth of the small intestine in humans. Crypt fission had an asymmetric peak mainly in neonates and infants, whereas crypt hyperplasia had a broad peak in infants and children, although both processes overlapped.

Our previous study of crypt fission and crypt hyperplasia in rats showed that crypt fission peaked early in infancy during milk feeding, whereas crypt hyperplasia peaked mainly after weaning, although the 2 processes overlapped ⁽¹³⁾. We have now shown further overlap of the 2 processes during infancy, but with evidence that crypt fission precedes crypt hyperplasia. This was because crypt fission was elevated in neonates over that of adults, whereas crypt hyperplasia was not present in neonates. Intestinal crypts form in utero in the intervillous spaces during intestinal development, and later crypt fission occurs prenatally ⁽³⁾. All of our infant subjects had mixed breast- or bottle feeding, so we were not able to separately examine the effect of breast-feeding on crypt fission. We previously distinguished subgroups of exclusively breast-fed and bottle-fed infants, and found that the crypt hyperplasia was not present in breast-fed but was present in bottle-fed infants ⁽⁴⁾. In that study, crypt fission was not examined. Because crypt fission occurs earlier than crypt hyperplasia and is associated with early intestinal organ growth, we presume it is relatively independent of the luminal environment. The process of crypt fission continues during infancy, and is supplemented by crypt hyperplasia that expands the surface area of the intestinal mucosa ⁽⁵⁾.

Crypt fission is believed to occur when the number of intestinal stem cells exceeds a 2-fold increased threshold ⁽⁷⁾. Stem cells divide symmetrically into 2 stem cells for this process. The Wnt/β-catenin pathway has been implicated in promoting intestinal stem cell division, although this largely is extrapolated from studies of aberrant crypt fission in models of juvenile polyposis and in humans and mice with mutations in the adenomatous polyposis coli gene ^(14,15,20). Crypt architecture is severely disrupted in genetically manipulated mice in which Wnt signalling is disrupted ^(21–25). There are 19 possible Wnt ligands, but which Wnt ligands are important physiologically remains uncertain ⁽²⁶⁾. We recently have shown that cellular expression of β-catenin peaks in cells of the stem cell region of intestinal crypts of the rat at a similar age to that of the peak of crypt fission ⁽²⁷⁾. Cytoplasmic and nuclear overexpression is the hallmark of activation of the β-catenin pathway. We also have some preliminary evidence that cellular expression of β-catenin also peaks in the stem cell region of intestinal crypts of human infants compared with adults (unpublished data, 2007). Taken together, these studies are compatible with the Wnt/β-catenin pathway being available to promote crypt fission.

The regulation of crypt hyperplasia is less well-defined. Our previous studies of crypt hyperplasia during infancy in both normal rats and humans showed that crypt hyperplasia was immunologically mediated and T cell dependent ^(28–32). In particular, we showed that there is a physiological peak of T cell activity during infancy in humans ⁽³⁰⁾, and that in rats this is localised to the gut-associated lymphoid tissue ⁽³²⁾. In rats, we have shown that crypt hyperplasia is inhibited by cyclosporin A ⁽²⁹⁾ or by anti–interleukin-2 receptor antibody ⁽³¹⁾. Activation of T cells in fetal intestinal explant culture is associated with crypt hyperplasia and upregulation of pericryptal keratinoctye growth factor expression ⁽³³⁾, but we have been unable to confirm any change in keratinocyte growth factor expression physiologically during growth of the small intestine in rats (unpublished data, 2005). Thus, no immunological growth factor or cytokine has been directly implicated in promoting crypt hyperplasia physiologically. Presumably, T cells activate a downstream nonimmune regulatory pathway. Crypt hyperplasia must arise from asymmetrical division of stem cells to a stem cell and a more mitotically active crypt progenitor cell. Either T cells could promote this pattern of stem cell division or else promote further division and expansion of the crypt progenitor cells.

Another regulatory pathway must therefore affect whether stem cells divide symmetrically to cause crypt fission or asymmetrically to cause crypt hyperplasia. A possible regulatory pathway is the Notch pathway ^(34,35). Activity of the Notch pathway maintains stem cells by promoting symmetrical cell division to 2 stem cells rather than a stem cell and a more differentiated progenitor cell that will form enteroendocrine cells, goblet cells, and Paneth cells ⁽³⁴⁾. Changes in activity of the Notch pathway also need to be studied with age. Other pathways that regulate stem cells are the bone morphogenetic protein and hedgehog pathways ^(14,34).

There are always inherent limitations with biopsies from humans compared with animals, given that all of our subjects had an independent medical indication for either surgery or endoscopy. It is obviously impossible to ethically obtain biopsies otherwise from healthy neonates, infants, or children. Although there is some regional decrease of villous length and crypt length along the small intestine from duodenum to jejunum to the ileum, we are confident that the surgical trimming in neonates could be compared with duodenal biopsies from infants, children, and adults. The large differences of villous area, crypt area, and crypt fission of neonates versus infants seen in our study were likely due to developmental age rather than regional differences. Older morphological studies have shown that neonates at birth have narrow fingerlike villi with small crypts, and that villi widen from fingerlike to leaflike and crypts become longer with age ^(36,37). Our study quantifies these age-dependent changes. The majority of our patients had reflux oesophagitis and some subjects would have been on histamine-2 antagonists or proton pump inhibitors. There is no evidence that these drugs change intestinal morphology or morphometry of the small intestine.

In conclusion, our study has shown that crypt fission peaks early in neonatal life and during infancy, subsequently decreases during late infancy and childhood, and is low in adults. Crypt hyperplasia was low in neonates and had a broad peak during infancy and childhood. Further studies are needed to investigate how intestinal stem cells are regulated to cause crypt fission or crypt hyperplasia.

Acknowledgment

The authors are grateful to Gary Goland for technical help with immunostaining of biopsies.